vpx_codec_dec_init: check that the iface is a decoder

Make sure the given interface is actually a decoder interface before initializing it. Change-Id: Ie48d737f2956cc2f0891666de5ea87251e96bc49
Remove unused vp8_get4x4sse_cs_mmx declaration
2011-03-24 15:05:10 +02:00 · 2011-03-24 15:05:10 +02:00 · 2011-03-24 15:05:10 +02:00 · 2011-03-24 15:05:10 +02:00 · 2011-03-24 15:05:09 +02:00 · 2011-03-24 15:05:09 +02:00
198 changed files with 4823 additions and 17888 deletions
--- a/.mailmap
+++ b/.mailmap
@@ -1,2 +1,4 @@
 Adrian Grange <agrange@google.com>
 Johann Koenig <johannkoenig@google.com>
+Tero Rintaluoma <teror@google.com> <tero.rintaluoma@on2.com>
+Tom Finegan <tomfinegan@google.com>
--- a/9
+++ b/9
@@ -4,13 +4,18 @@
 Aaron Watry <awatry@gmail.com>
 Adrian Grange <agrange@google.com>
 Alex Converse <alex.converse@gmail.com>
+Andoni Morales Alastruey <ylatuya@gmail.com>
 Andres Mejia <mcitadel@gmail.com>
+Attila Nagy <attilanagy@google.com>
 Fabio Pedretti <fabio.ped@libero.it>
 Frank Galligan <fgalligan@google.com>
 Fredrik Söderquist <fs@opera.com>
 Fritz Koenig <frkoenig@google.com>
+Gaute Strokkenes <gaute.strokkenes@broadcom.com>
 Giuseppe Scrivano <gscrivano@gnu.org>
 Guillermo Ballester Valor <gbvalor@gmail.com>
+Henrik Lundin <hlundin@google.com>
+James Berry <jamesberry@google.com>
 James Zern <jzern@google.com>
 Jan Kratochvil <jan.kratochvil@redhat.com>
 Jeff Muizelaar <jmuizelaar@mozilla.com>
@@ -23,10 +28,14 @@ Luca Barbato <lu_zero@gentoo.org>
 Makoto Kato <makoto.kt@gmail.com>
 Martin Ettl <ettl.martin78@googlemail.com>
 Michael Kohler <michaelkohler@live.com>
+Mikhal Shemer <mikhal@google.com>
+Pascal Massimino <pascal.massimino@gmail.com>
+Patrik Westin <patrik.westin@gmail.com>
 Paul Wilkins <paulwilkins@google.com>
 Pavol Rusnak <stick@gk2.sk>
 Philip Jägenstedt <philipj@opera.com>
 Scott LaVarnway <slavarnway@google.com>
+Tero Rintaluoma <teror@google.com>
 Timothy B. Terriberry <tterribe@xiph.org>
 Tom Finegan <tomfinegan@google.com>
 Yaowu Xu <yaowu@google.com>
--- a/77
+++ b/77
@@ -1,3 +1,80 @@
+2011-03-07 v0.9.6 "Bali"
+  Our second named release, focused on a faster, higher quality, encoder.
+
+  - Upgrading:
+    This release is backwards compatible with Aylesbury (v0.9.5). Users
+    of older releases should refer to the Upgrading notes in this
+    document for that release.
+
+  - Enhancements:
+      vpxenc --psnr shows a summary when encode completes
+      --tune=ssim option to enable activity masking
+      improved postproc visualizations for development
+      updated support for Apple iOS to SDK 4.2
+      query decoder to determine which reference frames were updated
+      implemented error tracking in the decoder
+      fix pipe support on windows
+
+  - Speed:
+      Primary focus was on good quality mode, speed 0. Average improvement
+      on x86 about 40%, up to 100% on user-generated content at that speed.
+      Best quality mode speed improved 35%, and realtime speed 10-20%. This
+      release also saw significant improvement in realtime encoding speed
+      on ARM platforms.
+
+        Improved encoder threading
+        Dont pick encoder filter level when loopfilter is disabled.
+        Avoid double copying of key frames into alt and golden buffer
+        FDCT optimizations.
+        x86 sse2 temporal filter
+        SSSE3 version of fast quantizer
+        vp8_rd_pick_best_mbsegmentation code restructure
+        Adjusted breakout RD for SPLITMV
+        Changed segmentation check order
+        Improved rd_pick_intra4x4block
+        Adds armv6 optimized variance calculation
+        ARMv6 optimized sad16x16
+        ARMv6 optimized half pixel variance calculations
+        Full search SAD function optimization in SSE4.1
+        Improve MV prediction accuracy to achieve performance gain
+        Improve MV prediction in vp8_pick_inter_mode() for speed>3
+
+  - Quality:
+      Best quality mode improved PSNR 6.3%, and SSIM 6.1%. This release
+      also includes support for "activity masking," which greatly improves
+      SSIM at the expense of PSNR. For now, this feature is available with
+      the --tune=ssim option. Further experimentation in this area
+      is ongoing. This release also introduces a new rate control mode
+      called "CQ," which changes the allocation of bits within a clip to
+      the sections where they will have the most visual impact.
+
+        Tuning for the more exact quantizer.
+        Relax rate control for last few frames
+        CQ Mode
+        Limit key frame quantizer for forced key frames.
+        KF/GF Pulsing
+        Add simple version of activity masking.
+        make rdmult adaptive for intra in quantizer RDO
+        cap the best quantizer for 2nd order DC
+        change the threshold of DC check for encode breakout
+
+  - Bug Fixes:
+      Fix crash on Sparc Solaris.
+      Fix counter of fixed keyframe distance
+      ARNR filter pointer update bug fix
+      Fixed use of motion percentage in KF/GF group calc
+      Changed condition for using RD in Intra Mode
+      Fix encoder real-time only configuration.
+      Fix ARM encoder crash with multiple token partitions
+      Fixed bug first cluster timecode of webm file is wrong.
+      Fixed various encoder bugs with odd-sized images
+      vp8e_get_preview fixed when spatial resampling enabled
+      quantizer: fix assertion in fast quantizer path
+      Allocate source buffers to be multiples of 16
+      Fix for manual Golden frame frequency
+      Fix drastic undershoot in long form content
+
+
 2010-10-28 v0.9.5 "Aylesbury"
  Our first named release, focused on a faster decoder, and a better encoder.

--- a/4
+++ b/4
@@ -45,18 +45,14 @@ COMPILING THE APPLICATIONS/LIBRARIES:
    armv5te-linux-rvct
    armv5te-linux-gcc
    armv5te-symbian-gcc
-    armv5te-wince-vs8
    armv6-darwin-gcc
    armv6-linux-rvct
    armv6-linux-gcc
    armv6-symbian-gcc
-    armv6-wince-vs8
    iwmmxt-linux-rvct
    iwmmxt-linux-gcc
-    iwmmxt-wince-vs8
    iwmmxt2-linux-rvct
    iwmmxt2-linux-gcc
-    iwmmxt2-wince-vs8
    armv7-linux-rvct
    armv7-linux-gcc
    mips32-linux-gcc
--- a/build/arm-wince-vs8/armasmv5.rules
+++ b/build/arm-wince-vs8/armasmv5.rules
@@ -1,20 +0,0 @@
-<?xml version="1.0" encoding="utf-8"?>
-<VisualStudioToolFile
-	Name="armasm"
-	Version="8.00"
-	>
-	<Rules>
-		<CustomBuildRule
-			Name="ARMASM"
-			DisplayName="Armasm Assembler"
-			CommandLine="armasm -o &quot;$(IntDir)\$(InputName).obj&quot; $(InputPath) -32 -ARCH 5&#x0D;&#x0A;"
-			Outputs="$(IntDir)\$(InputName).obj"
-			FileExtensions="*.asm"
-			ExecutionDescription="Assembling $(InputName).asm"
-			ShowOnlyRuleProperties="false"
-			>
-			<Properties>
-			</Properties>
-		</CustomBuildRule>
-	</Rules>
-</VisualStudioToolFile>
--- a/build/arm-wince-vs8/armasmv6.rules
+++ b/build/arm-wince-vs8/armasmv6.rules
@@ -1,20 +0,0 @@
-<?xml version="1.0" encoding="utf-8"?>
-<VisualStudioToolFile
-	Name="armasm"
-	Version="8.00"
-	>
-	<Rules>
-		<CustomBuildRule
-			Name="ARMASM"
-			DisplayName="Armasm Assembler"
-			CommandLine="armasm -o &quot;$(IntDir)\$(InputName).obj&quot; $(InputPath) -32 -ARCH 6&#x0D;&#x0A;"
-			Outputs="$(IntDir)\$(InputName).obj"
-			FileExtensions="*.asm"
-			ExecutionDescription="Assembling $(InputName).asm"
-			ShowOnlyRuleProperties="false"
-			>
-			<Properties>
-			</Properties>
-		</CustomBuildRule>
-	</Rules>
-</VisualStudioToolFile>
--- a/build/arm-wince-vs8/armasmxscale.rules
+++ b/build/arm-wince-vs8/armasmxscale.rules
@@ -1,20 +0,0 @@
-<?xml version="1.0" encoding="utf-8"?>
-<VisualStudioToolFile
-	Name="armasm"
-	Version="8.00"
-	>
-	<Rules>
-		<CustomBuildRule
-			Name="ARMASM"
-			DisplayName="Armasm Assembler"
-			CommandLine="armasm -o &quot;$(IntDir)\$(InputName).obj&quot; $(InputPath) -32 -cpu XSCALE&#x0D;&#x0A;"
-			Outputs="$(IntDir)\$(InputName).obj"
-			FileExtensions="*.asm"
-			ExecutionDescription="Assembling $(InputName).asm"
-			ShowOnlyRuleProperties="false"
-			>
-			<Properties>
-			</Properties>
-		</CustomBuildRule>
-	</Rules>
-</VisualStudioToolFile>
--- a/build/arm-wince-vs8/obj_int_extract.bat
+++ b/build/arm-wince-vs8/obj_int_extract.bat
@@ -1,13 +0,0 @@
-@echo off
-REM   Copyright (c) 2010 The WebM project authors. All Rights Reserved.
-REM
-REM   Use of this source code is governed by a BSD-style license
-REM   that can be found in the LICENSE file in the root of the source
-REM   tree. An additional intellectual property rights grant can be found
-REM   in the file PATENTS.  All contributing project authors may
-REM   be found in the AUTHORS file in the root of the source tree.
-echo on
-
-
-cl /I ".\\" /I "..\vp6_decoder_sdk" /I "..\vp6_decoder_sdk\vpx_ports" /D "NDEBUG" /D "_WIN32_WCE=0x420" /D "UNDER_CE" /D "WIN32_PLATFORM_PSPC" /D "WINCE" /D "_LIB" /D "ARM" /D "_ARM_" /D "_UNICODE" /D "UNICODE" /FD /EHsc /MT /GS- /fp:fast /GR- /Fo"Pocket_PC_2003__ARMV4_\%1/" /Fd"Pocket_PC_2003__ARMV4_\%1/vc80.pdb" /W3 /nologo /c /TC ..\vp6_decoder_sdk\vp6_decoder\algo\common\arm\dec_asm_offsets_arm.c
-obj_int_extract.exe rvds "Pocket_PC_2003__ARMV4_\%1/dec_asm_offsets_arm.obj"
--- a/build/arm-wince-vs8/vpx.sln
+++ b/build/arm-wince-vs8/vpx.sln
@@ -1,88 +0,0 @@
-Microsoft Visual Studio Solution File, Format Version 9.00
-# Visual Studio 2005
-Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "example", "example.vcproj", "{BA5FE66F-38DD-E034-F542-B1578C5FB950}"
-	ProjectSection(ProjectDependencies) = postProject
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74} = {DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2} = {E1360C65-D375-4335-8057-7ED99CC3F9B2}
-	EndProjectSection
-EndProject
-Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "obj_int_extract", "obj_int_extract.vcproj", "{E1360C65-D375-4335-8057-7ED99CC3F9B2}"
-EndProject
-Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "vpx", "vpx.vcproj", "{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}"
-	ProjectSection(ProjectDependencies) = postProject
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2} = {E1360C65-D375-4335-8057-7ED99CC3F9B2}
-	EndProjectSection
-EndProject
-Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "xma", "xma.vcproj", "{A955FC4A-73F1-44F7-135E-30D84D32F022}"
-	ProjectSection(ProjectDependencies) = postProject
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2} = {E1360C65-D375-4335-8057-7ED99CC3F9B2}
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74} = {DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}
-	EndProjectSection
-EndProject
-Global
-	GlobalSection(SolutionConfigurationPlatforms) = preSolution
-		Debug|Mixed Platforms = Debug|Mixed Platforms
-		Debug|Pocket PC 2003 (ARMV4) = Debug|Pocket PC 2003 (ARMV4)
-		Debug|Win32 = Debug|Win32
-		Release|Mixed Platforms = Release|Mixed Platforms
-		Release|Pocket PC 2003 (ARMV4) = Release|Pocket PC 2003 (ARMV4)
-		Release|Win32 = Release|Win32
-	EndGlobalSection
-	GlobalSection(ProjectConfigurationPlatforms) = postSolution
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Mixed Platforms.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Mixed Platforms.Build.0 = Debug|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Mixed Platforms.Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Pocket PC 2003 (ARMV4).ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Pocket PC 2003 (ARMV4).Build.0 = Debug|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Pocket PC 2003 (ARMV4).Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Win32.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Mixed Platforms.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Mixed Platforms.Build.0 = Release|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Mixed Platforms.Deploy.0 = Release|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Pocket PC 2003 (ARMV4).Build.0 = Release|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Pocket PC 2003 (ARMV4).Deploy.0 = Release|Pocket PC 2003 (ARMV4)
-		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Win32.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Mixed Platforms.ActiveCfg = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Mixed Platforms.Build.0 = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Win32.ActiveCfg = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Win32.Build.0 = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Mixed Platforms.ActiveCfg = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Mixed Platforms.Build.0 = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Win32.ActiveCfg = Release|Win32
-		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Win32.Build.0 = Release|Win32
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Mixed Platforms.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Mixed Platforms.Build.0 = Debug|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Mixed Platforms.Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Pocket PC 2003 (ARMV4).ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Pocket PC 2003 (ARMV4).Build.0 = Debug|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Pocket PC 2003 (ARMV4).Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Win32.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Mixed Platforms.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Mixed Platforms.Build.0 = Release|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Mixed Platforms.Deploy.0 = Release|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Pocket PC 2003 (ARMV4).Build.0 = Release|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Pocket PC 2003 (ARMV4).Deploy.0 = Release|Pocket PC 2003 (ARMV4)
-		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Win32.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Mixed Platforms.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Mixed Platforms.Build.0 = Debug|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Mixed Platforms.Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Pocket PC 2003 (ARMV4).ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Pocket PC 2003 (ARMV4).Build.0 = Debug|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Pocket PC 2003 (ARMV4).Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Win32.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Mixed Platforms.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Mixed Platforms.Build.0 = Release|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Mixed Platforms.Deploy.0 = Release|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Pocket PC 2003 (ARMV4).Build.0 = Release|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Pocket PC 2003 (ARMV4).Deploy.0 = Release|Pocket PC 2003 (ARMV4)
-		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Win32.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
-	EndGlobalSection
-	GlobalSection(SolutionProperties) = preSolution
-		HideSolutionNode = FALSE
-	EndGlobalSection
-EndGlobal
--- a/build/make/Makefile
+++ b/build/make/Makefile
@@ -152,8 +152,8 @@ endif
 # Rule to extract assembly constants from C sources
 #
 obj_int_extract: build/make/obj_int_extract.c
-	$(if $(quiet),echo "    [HOSTCC] $@")
-	$(qexec)$(HOSTCC) -I. -o $@ $<
+	$(if $(quiet),@echo "    [HOSTCC] $@")
+	$(qexec)$(HOSTCC) -I. -I$(SRC_PATH_BARE) -o $@ $<
 CLEAN-OBJS += obj_int_extract

 #
@@ -255,7 +255,7 @@ ifeq ($(filter clean,$(MAKECMDGOALS)),)
 endif

 #
-# Configuration dependant rules
+# Configuration dependent rules
 #
 $(call pairmap,install_map_templates,$(INSTALL_MAPS))

@@ -332,7 +332,7 @@ ifneq ($(call enabled,DIST-SRCS),)
    DIST-SRCS-$(CONFIG_MSVS)  += build/x86-msvs/yasm.rules
    DIST-SRCS-$(CONFIG_RVCT) += build/make/armlink_adapter.sh
    #
-    # This isn't really ARCH_ARM dependent, it's dependant on whether we're
+    # This isn't really ARCH_ARM dependent, it's dependent on whether we're
    # using assembly code or not (CONFIG_OPTIMIZATIONS maybe). Just use
    # this for now.
    DIST-SRCS-$(ARCH_ARM)    += build/make/obj_int_extract.c
--- a/build/make/configure.sh
+++ b/build/make/configure.sh
@@ -83,7 +83,7 @@ Build options:
  ${toggle_werror}            treat warnings as errors, if possible
                              (not available with all compilers)
  ${toggle_optimizations}     turn on/off compiler optimization flags
-  ${toggle_pic}               turn on/off Position Independant Code
+  ${toggle_pic}               turn on/off Position Independent Code
  ${toggle_ccache}            turn on/off compiler cache
  ${toggle_debug}             enable/disable debug mode
  ${toggle_gprof}             enable/disable gprof profiling instrumentation
@@ -624,6 +624,10 @@ process_common_toolchain() {

    # Handle Solaris variants. Solaris 10 needs -lposix4
    case ${toolchain} in
+        sparc-solaris-*)
+            add_extralibs -lposix4
+            add_cflags "-DMUST_BE_ALIGNED"
+            ;;
        *-solaris-*)
            add_extralibs -lposix4
            ;;
@@ -863,7 +867,7 @@ process_common_toolchain() {
                setup_gnu_toolchain
                add_cflags -use-msasm -use-asm
                add_ldflags -i-static
-                enabled x86_64 && add_cflags -ipo -no-prec-div -static -xSSE3 -axSSE3
+                enabled x86_64 && add_cflags -ipo -no-prec-div -static -xSSE2 -axSSE2
                enabled x86_64 && AR=xiar
                case ${tune_cpu} in
                    atom*)
@@ -953,7 +957,7 @@ process_common_toolchain() {
        enabled small && check_add_cflags -O2 || check_add_cflags -O3
    fi

-    # Position Independant Code (PIC) support, for building relocatable
+    # Position Independent Code (PIC) support, for building relocatable
    # shared objects
    enabled gcc && enabled pic && check_add_cflags -fPIC

--- a/build/make/gen_msvs_proj.sh
+++ b/build/make/gen_msvs_proj.sh
@@ -32,7 +32,8 @@ Options:
    --name=project_name         Name of the project (required)
    --proj-guid=GUID            GUID to use for the project
    --module-def=filename       File containing export definitions (for DLLs)
-    --ver=version               Version (7,8) of visual studio to generate for
+    --ver=version               Version (7,8,9) of visual studio to generate for
+    --src-path-bare=dir         Path to root of source tree
    -Ipath/to/include           Additional include directories
    -DFLAG[=value]              Preprocessor macros to define
    -Lpath/to/lib               Additional library search paths
@@ -132,7 +133,7 @@ generate_filter() {
    open_tag Filter \
        Name=$name \
        Filter=$pats \
-        UniqueIdentifier=`generate_uuid`
+        UniqueIdentifier=`generate_uuid` \

    file_list_sz=${#file_list[@]}
    for i in ${!file_list[@]}; do
@@ -145,31 +146,21 @@ generate_filter() {
                if [ "$pat" == "asm" ] && $asm_use_custom_step; then
                    for plat in "${platforms[@]}"; do
                        for cfg in Debug Release; do
-                            open_tag  FileConfiguration \
-                            Name="${cfg}|${plat}"
+                            open_tag FileConfiguration \
+                                Name="${cfg}|${plat}" \
+
                            tag Tool \
                                Name="VCCustomBuildTool" \
                                Description="Assembling \$(InputFileName)" \
-                                CommandLine="$(eval echo \$asm_${cfg}_cmdline)"\
-                                Outputs="\$(InputName).obj"
+                                CommandLine="$(eval echo \$asm_${cfg}_cmdline)" \
+                                Outputs="\$(InputName).obj" \
+
                            close_tag FileConfiguration
                        done
                    done
                fi

-                if [ "${f##*.}" == "cpp" ]; then
-                    for plat in "${platforms[@]}"; do
-                        for cfg in Debug Release; do
-                        open_tag FileConfiguration \
-                            Name="${cfg}|${plat}"
-                        tag Tool \
-                            Name="VCCLCompilerTool" \
-                            CompileAs="2"
-                        close_tag FileConfiguration
-                        done
-                    done
-                fi
-                close_tag  File
+                close_tag File

                break
            fi
@@ -185,57 +176,63 @@ unset target
 for opt in "$@"; do
    optval="${opt#*=}"
    case "$opt" in
-    --help|-h) show_help
-    ;;
-    --target=*) target="${optval}"
-    ;;
-    --out=*) outfile="$optval"
-    ;;
-    --name=*) name="${optval}"
-    ;;
-    --proj-guid=*) guid="${optval}"
-    ;;
-    --module-def=*)
-        link_opts="${link_opts} ModuleDefinitionFile=${optval}"
-    ;;
-    --exe) proj_kind="exe"
-    ;;
-    --lib) proj_kind="lib"
-    ;;
-    --static-crt) use_static_runtime=true
-    ;;
-    --ver=*) vs_ver="$optval"
-             case $optval in
-             [789])
-             ;;
-             *) die Unrecognized Visual Studio Version in $opt
-             ;;
-             esac
-    ;;
-    -I*) opt="${opt%/}"
-         incs="${incs}${incs:+;}&quot;${opt##-I}&quot;"
-         yasmincs="${yasmincs} ${opt}"
-    ;;
-    -D*) defines="${defines}${defines:+;}${opt##-D}"
-    ;;
-    -L*) # fudge . to $(OutDir)
-         if [ "${opt##-L}" == "." ]; then
-             libdirs="${libdirs}${libdirs:+;}&quot;\$(OutDir)&quot;"
-         else
-             # Also try directories for this platform/configuration
-             libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}&quot;"
-             libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}/\$(PlatformName)/\$(ConfigurationName)&quot;"
-             libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}/\$(PlatformName)&quot;"
-         fi
-    ;;
-    -l*) libs="${libs}${libs:+ }${opt##-l}.lib"
-    ;;
-    -*) die_unknown $opt
-    ;;
-    *) file_list[${#file_list[@]}]="$opt"
-       case "$opt" in
-       *.asm) uses_asm=true;;
-       esac
+        --help|-h) show_help
+        ;;
+        --target=*) target="${optval}"
+        ;;
+        --out=*) outfile="$optval"
+        ;;
+        --name=*) name="${optval}"
+        ;;
+        --proj-guid=*) guid="${optval}"
+        ;;
+        --module-def=*) link_opts="${link_opts} ModuleDefinitionFile=${optval}"
+        ;;
+        --exe) proj_kind="exe"
+        ;;
+        --lib) proj_kind="lib"
+        ;;
+        --src-path-bare=*) src_path_bare="$optval"
+        ;;
+        --static-crt) use_static_runtime=true
+        ;;
+        --ver=*)
+            vs_ver="$optval"
+            case "$optval" in
+                [789])
+                ;;
+                *) die Unrecognized Visual Studio Version in $opt
+                ;;
+            esac
+        ;;
+        -I*)
+            opt="${opt%/}"
+            incs="${incs}${incs:+;}&quot;${opt##-I}&quot;"
+            yasmincs="${yasmincs} ${opt}"
+        ;;
+        -D*) defines="${defines}${defines:+;}${opt##-D}"
+        ;;
+        -L*) # fudge . to $(OutDir)
+            if [ "${opt##-L}" == "." ]; then
+                libdirs="${libdirs}${libdirs:+;}&quot;\$(OutDir)&quot;"
+            else
+                 # Also try directories for this platform/configuration
+                 libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}&quot;"
+                 libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}/\$(PlatformName)/\$(ConfigurationName)&quot;"
+                 libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}/\$(PlatformName)&quot;"
+            fi
+        ;;
+        -l*) libs="${libs}${libs:+ }${opt##-l}.lib"
+        ;;
+        -*) die_unknown $opt
+        ;;
+        *)
+            file_list[${#file_list[@]}]="$opt"
+            case "$opt" in
+                 *.asm) uses_asm=true
+                 ;;
+            esac
+        ;;
    esac
 done
 outfile=${outfile:-/dev/stdout}
@@ -278,11 +275,7 @@ done

 # List Keyword for this target
 case "$target" in
-    x86*)
-        keyword="ManagedCProj"
-    ;;
-    arm*|iwmmx*)
-        keyword="Win32Proj"
+    x86*) keyword="ManagedCProj"
    ;;
    *) die "Unsupported target $target!"
 esac
@@ -298,402 +291,255 @@ case "$target" in
        asm_Debug_cmdline="yasm -Xvc -g cv8 -f \$(PlatformName) ${yasmincs} &quot;\$(InputPath)&quot;"
        asm_Release_cmdline="yasm -Xvc -f \$(PlatformName) ${yasmincs} &quot;\$(InputPath)&quot;"
    ;;
-    arm*|iwmmx*)
-        case "${name}" in
-        obj_int_extract) platforms[0]="Win32"
-        ;;
-        *) platforms[0]="Pocket PC 2003 (ARMV4)"
-        ;;
-        esac
-    ;;
    *) die "Unsupported target $target!"
-esac
-
-# List Command-line Arguments for this target
-case "$target" in
-    arm*|iwmmx*)
-        if [ "$name" == "example" ];then
-            ARGU="--codec vp6 --flipuv --progress _bnd.vp6"
-        fi
-        if [ "$name" == "xma" ];then
-            ARGU="--codec vp6 -h 240 -w 320 -v"
-        fi
    ;;
 esac

 generate_vcproj() {
    case "$proj_kind" in
-    exe) vs_ConfigurationType=1
-    ;;
-    *)   vs_ConfigurationType=4
-    ;;
+        exe) vs_ConfigurationType=1
+        ;;
+        *)   vs_ConfigurationType=4
+        ;;
    esac

    echo "<?xml version=\"1.0\" encoding=\"Windows-1252\"?>"
-    open_tag  VisualStudioProject \
-                  ProjectType="Visual C++" \
-                  Version="${vs_ver_id}" \
-                  Name="${name}" \
-                  ProjectGUID="{${guid}}" \
-                  RootNamespace="${name}" \
-                  Keyword="${keyword}"
+    open_tag VisualStudioProject \
+        ProjectType="Visual C++" \
+        Version="${vs_ver_id}" \
+        Name="${name}" \
+        ProjectGUID="{${guid}}" \
+        RootNamespace="${name}" \
+        Keyword="${keyword}" \

-    open_tag  Platforms
+    open_tag Platforms
    for plat in "${platforms[@]}"; do
-        tag   Platform Name="$plat"
+        tag Platform Name="$plat"
    done
    close_tag Platforms

-    open_tag  ToolFiles
+    open_tag ToolFiles
    case "$target" in
        x86*) $uses_asm && tag ToolFile RelativePath="$self_dirname/../x86-msvs/yasm.rules"
        ;;
-        arm*|iwmmx*)
-            if [ "$name" == "vpx" ];then
-            case "$target" in
-                armv5*)
-                    tag ToolFile RelativePath="$self_dirname/../arm-wince-vs8/armasmv5.rules"
-                ;;
-                armv6*)
-                    tag ToolFile RelativePath="$self_dirname/../arm-wince-vs8/armasmv6.rules"
-                ;;
-                iwmmxt*)
-                    tag ToolFile RelativePath="$self_dirname/../arm-wince-vs8/armasmxscale.rules"
-                ;;
-            esac
-            fi
-        ;;
    esac
    close_tag ToolFiles

-    open_tag  Configurations
+    open_tag Configurations
    for plat in "${platforms[@]}"; do
        plat_no_ws=`echo $plat | sed 's/[^A-Za-z0-9_]/_/g'`
-        open_tag  Configuration \
-                      Name="Debug|$plat" \
-                      OutputDirectory="\$(SolutionDir)$plat_no_ws/\$(ConfigurationName)" \
-                      IntermediateDirectory="$plat_no_ws/\$(ConfigurationName)/${name}" \
-                      ConfigurationType="$vs_ConfigurationType" \
-                      CharacterSet="1"
-
-        if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
-            case "$name" in
-                vpx)         tag Tool \
-                             Name="VCPreBuildEventTool" \
-                             CommandLine="call obj_int_extract.bat \$(ConfigurationName)"
-                             tag Tool \
-                             Name="VCMIDLTool" \
-                             TargetEnvironment="1"
-                             tag Tool \
-                             Name="VCCLCompilerTool" \
-                             ExecutionBucket="7" \
-                             Optimization="0" \
-                             AdditionalIncludeDirectories="$incs" \
-                             PreprocessorDefinitions="_DEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES);WINCE;DEBUG;_LIB;\$(ARCHFAM);\$(_ARCHFAM_);_UNICODE;UNICODE;" \
-                             MinimalRebuild="true" \
-                             RuntimeLibrary="1" \
-                             BufferSecurityCheck="false" \
-                             UsePrecompiledHeader="0" \
-                             WarningLevel="3" \
-                             DebugInformationFormat="1" \
-                             CompileAs="1"
-                             tag Tool \
-                             Name="VCResourceCompilerTool" \
-                             PreprocessorDefinitions="_DEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES)" \
-                             Culture="1033" \
-                             AdditionalIncludeDirectories="\$(IntDir)" \
-                ;;
-                example|xma) tag Tool \
-                             Name="VCCLCompilerTool" \
-                             ExecutionBucket="7" \
-                             Optimization="0" \
-                             AdditionalIncludeDirectories="$incs" \
-                             PreprocessorDefinitions="_DEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES);WINCE;DEBUG;_CONSOLE;\$(ARCHFAM);\$(_ARCHFAM_);_UNICODE;UNICODE;" \
-                             MinimalRebuild="true" \
-                             RuntimeLibrary="1" \
-                             BufferSecurityCheck="false" \
-                             UsePrecompiledHeader="0" \
-                             WarningLevel="3" \
-                             DebugInformationFormat="1" \
-                             CompileAs="1"
-                             tag Tool \
-                             Name="VCResourceCompilerTool" \
-                             PreprocessorDefinitions="_DEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES)" \
-                             Culture="1033" \
-                             AdditionalIncludeDirectories="\$(IntDir)" \
-                ;;
-                obj_int_extract) tag Tool \
-                             Name="VCCLCompilerTool" \
-                             Optimization="0" \
-                             AdditionalIncludeDirectories="$incs" \
-                             PreprocessorDefinitions="WIN32;DEBUG;_CONSOLE" \
-                             RuntimeLibrary="1" \
-                             WarningLevel="3" \
-                             DebugInformationFormat="1" \
-                ;;
-            esac
-        fi
+        open_tag Configuration \
+            Name="Debug|$plat" \
+            OutputDirectory="\$(SolutionDir)$plat_no_ws/\$(ConfigurationName)" \
+            IntermediateDirectory="$plat_no_ws/\$(ConfigurationName)/${name}" \
+            ConfigurationType="$vs_ConfigurationType" \
+            CharacterSet="1" \

        case "$target" in
-            x86*) tag Tool \
-                Name="VCCLCompilerTool" \
-                Optimization="0" \
-                AdditionalIncludeDirectories="$incs" \
-                PreprocessorDefinitions="WIN32;_DEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
-                RuntimeLibrary="$debug_runtime" \
-                UsePrecompiledHeader="0" \
-                WarningLevel="3" \
-                DebugInformationFormat="1" \
-                Detect64BitPortabilityProblems="true" \
+            x86*)
+                case "$name" in
+                    obj_int_extract)
+                        tag Tool \
+                            Name="VCCLCompilerTool" \
+                            Optimization="0" \
+                            AdditionalIncludeDirectories="$incs" \
+                            PreprocessorDefinitions="WIN32;DEBUG;_CONSOLE;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE" \
+                            RuntimeLibrary="$debug_runtime" \
+                            WarningLevel="3" \
+                            Detect64BitPortabilityProblems="true" \
+                            DebugInformationFormat="1" \
+                    ;;
+                    vpx)
+                        tag Tool \
+                            Name="VCPreBuildEventTool" \
+                            CommandLine="call obj_int_extract.bat $src_path_bare" \

-                $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs" Debug="1"
+                        tag Tool \
+                            Name="VCCLCompilerTool" \
+                            Optimization="0" \
+                            AdditionalIncludeDirectories="$incs" \
+                            PreprocessorDefinitions="WIN32;_DEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
+                            RuntimeLibrary="$debug_runtime" \
+                            UsePrecompiledHeader="0" \
+                            WarningLevel="3" \
+                            DebugInformationFormat="1" \
+                            Detect64BitPortabilityProblems="true" \
+
+                        $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs" Debug="1"
+                    ;;
+                    *)
+                        tag Tool \
+                            Name="VCCLCompilerTool" \
+                            Optimization="0" \
+                            AdditionalIncludeDirectories="$incs" \
+                            PreprocessorDefinitions="WIN32;_DEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
+                            RuntimeLibrary="$debug_runtime" \
+                            UsePrecompiledHeader="0" \
+                            WarningLevel="3" \
+                            DebugInformationFormat="1" \
+                            Detect64BitPortabilityProblems="true" \
+
+                        $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs" Debug="1"
+                    ;;
+                esac
            ;;
        esac

        case "$proj_kind" in
            exe)
                case "$target" in
-                    x86*) tag Tool \
-                          Name="VCLinkerTool" \
-                          AdditionalDependencies="$debug_libs \$(NoInherit)" \
-                          AdditionalLibraryDirectories="$libdirs" \
-                          GenerateDebugInformation="true" \
-                          ProgramDatabaseFile="\$(OutDir)/${name}.pdb" \
-
-                    ;;
-                    arm*|iwmmx*)
+                    x86*)
                        case "$name" in
-                            obj_int_extract) tag Tool \
-                                Name="VCLinkerTool" \
-                                OutputFile="${name}.exe" \
-                                GenerateDebugInformation="true"
+                            obj_int_extract)
+                                tag Tool \
+                                    Name="VCLinkerTool" \
+                                    OutputFile="${name}.exe" \
+                                    GenerateDebugInformation="true" \
                            ;;
-                            *) tag Tool \
-                                Name="VCLinkerTool" \
-                                AdditionalDependencies="$debug_libs" \
-                                OutputFile="\$(OutDir)/${name}.exe" \
-                                LinkIncremental="2" \
-                                AdditionalLibraryDirectories="${libdirs};&quot;..\lib/$plat_no_ws&quot;" \
-                                DelayLoadDLLs="\$(NOINHERIT)" \
-                                GenerateDebugInformation="true" \
-                                ProgramDatabaseFile="\$(OutDir)/${name}.pdb" \
-                                SubSystem="9" \
-                                StackReserveSize="65536" \
-                                StackCommitSize="4096" \
-                                EntryPointSymbol="mainWCRTStartup" \
-                                TargetMachine="3"
+                            *)
+                                tag Tool \
+                                    Name="VCLinkerTool" \
+                                    AdditionalDependencies="$debug_libs \$(NoInherit)" \
+                                    AdditionalLibraryDirectories="$libdirs" \
+                                    GenerateDebugInformation="true" \
+                                    ProgramDatabaseFile="\$(OutDir)/${name}.pdb" \
                            ;;
                        esac
-                     ;;
+                    ;;
                 esac
            ;;
            lib)
                case "$target" in
-                      arm*|iwmmx*) tag Tool \
-                                    Name="VCLibrarianTool" \
-                                    AdditionalOptions=" /subsystem:windowsce,4.20 /machine:ARM" \
-                                    OutputFile="\$(OutDir)/${name}.lib" \
-                                ;;
-                                *) tag Tool \
-                                    Name="VCLibrarianTool" \
-                                    OutputFile="\$(OutDir)/${name}${lib_sfx}d.lib" \
-                                ;;
+                    x86*)
+                        tag Tool \
+                            Name="VCLibrarianTool" \
+                            OutputFile="\$(OutDir)/${name}${lib_sfx}d.lib" \
+
+                    ;;
                esac
            ;;
-            dll) tag Tool \
-                 Name="VCLinkerTool" \
-                AdditionalDependencies="\$(NoInherit)" \
-                LinkIncremental="2" \
-                GenerateDebugInformation="true" \
-                AssemblyDebug="1" \
-                TargetMachine="1" \
-                      $link_opts
+            dll)
+                tag Tool \
+                    Name="VCLinkerTool" \
+                    AdditionalDependencies="\$(NoInherit)" \
+                    LinkIncremental="2" \
+                    GenerateDebugInformation="true" \
+                    AssemblyDebug="1" \
+                    TargetMachine="1" \
+                    $link_opts \
+
+            ;;
        esac

-        if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
-            case "$name" in
-                vpx)         tag DeploymentTool \
-                             ForceDirty="-1" \
-                             RegisterOutput="0"
-                                ;;
-                example|xma) tag DeploymentTool \
-                             ForceDirty="-1" \
-                             RegisterOutput="0"
-                             tag DebuggerTool \
-                             Arguments="${ARGU}"
-                                ;;
-            esac
-        fi
        close_tag Configuration

-        open_tag  Configuration \
-                      Name="Release|$plat" \
-                      OutputDirectory="\$(SolutionDir)$plat_no_ws/\$(ConfigurationName)" \
-                      IntermediateDirectory="$plat_no_ws/\$(ConfigurationName)/${name}" \
-                      ConfigurationType="$vs_ConfigurationType" \
-                      CharacterSet="1" \
-                      WholeProgramOptimization="0"
+        open_tag Configuration \
+            Name="Release|$plat" \
+            OutputDirectory="\$(SolutionDir)$plat_no_ws/\$(ConfigurationName)" \
+            IntermediateDirectory="$plat_no_ws/\$(ConfigurationName)/${name}" \
+            ConfigurationType="$vs_ConfigurationType" \
+            CharacterSet="1" \
+            WholeProgramOptimization="0" \

-        if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
-            case "$name" in
-                vpx)         tag Tool \
-                                     Name="VCPreBuildEventTool" \
-                                     CommandLine="call obj_int_extract.bat \$(ConfigurationName)"
-                             tag Tool \
-                                     Name="VCMIDLTool" \
-                                     TargetEnvironment="1"
-                             tag Tool \
-                                             Name="VCCLCompilerTool" \
-                                             ExecutionBucket="7" \
-                                             Optimization="2" \
-                                             FavorSizeOrSpeed="1" \
-                                             AdditionalIncludeDirectories="$incs" \
-                                             PreprocessorDefinitions="NDEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES);WINCE;_LIB;\$(ARCHFAM);\$(_ARCHFAM_);_UNICODE;UNICODE;" \
-                                             RuntimeLibrary="0" \
-                                             BufferSecurityCheck="false" \
-                                             UsePrecompiledHeader="0" \
-                                             WarningLevel="3" \
-                                             DebugInformationFormat="0" \
-                                             CompileAs="1"
-                             tag Tool \
-                                             Name="VCResourceCompilerTool" \
-                                             PreprocessorDefinitions="NDEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES)" \
-                                             Culture="1033" \
-                                             AdditionalIncludeDirectories="\$(IntDir)" \
-                ;;
-                example|xma) tag Tool \
-                             Name="VCCLCompilerTool" \
-                             ExecutionBucket="7" \
-                             Optimization="2" \
-                             FavorSizeOrSpeed="1" \
-                             AdditionalIncludeDirectories="$incs" \
-                             PreprocessorDefinitions="NDEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES);WINCE;_CONSOLE;\$(ARCHFAM);\$(_ARCHFAM_);_UNICODE;UNICODE;" \
-                             RuntimeLibrary="0" \
-                             BufferSecurityCheck="false" \
-                             UsePrecompiledHeader="0" \
-                             WarningLevel="3" \
-                             DebugInformationFormat="0" \
-                             CompileAs="1"
-                             tag Tool \
-                             Name="VCResourceCompilerTool" \
-                             PreprocessorDefinitions="NDEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES)" \
-                             Culture="1033" \
-                             AdditionalIncludeDirectories="\$(IntDir)" \
-                ;;
-                obj_int_extract) tag Tool \
-                             Name="VCCLCompilerTool" \
-                             AdditionalIncludeDirectories="$incs" \
-                             PreprocessorDefinitions="WIN32;NDEBUG;_CONSOLE" \
-                             RuntimeLibrary="0" \
-                             UsePrecompiledHeader="0" \
-                             WarningLevel="3" \
-                             Detect64BitPortabilityProblems="true" \
-                             DebugInformationFormat="0" \
-                ;;
-            esac
-        fi
+        case "$target" in
+            x86*)
+                case "$name" in
+                    obj_int_extract)
+                        tag Tool \
+                            Name="VCCLCompilerTool" \
+                            AdditionalIncludeDirectories="$incs" \
+                            PreprocessorDefinitions="WIN32;NDEBUG;_CONSOLE;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE" \
+                            RuntimeLibrary="$release_runtime" \
+                            UsePrecompiledHeader="0" \
+                            WarningLevel="3" \
+                            Detect64BitPortabilityProblems="true" \
+                            DebugInformationFormat="0" \
+                    ;;
+                    vpx)
+                        tag Tool \
+                            Name="VCPreBuildEventTool" \
+                            CommandLine="call obj_int_extract.bat $src_path_bare" \

-    case "$target" in
-        x86*) tag       Tool \
-                      Name="VCCLCompilerTool" \
-                      AdditionalIncludeDirectories="$incs" \
-                      PreprocessorDefinitions="WIN32;NDEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
-                      RuntimeLibrary="$release_runtime" \
-                      UsePrecompiledHeader="0" \
-                      WarningLevel="3" \
-                      DebugInformationFormat="0" \
-                      Detect64BitPortabilityProblems="true"
+                        tag Tool \
+                            Name="VCCLCompilerTool" \
+                            AdditionalIncludeDirectories="$incs" \
+                            PreprocessorDefinitions="WIN32;NDEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
+                            RuntimeLibrary="$release_runtime" \
+                            UsePrecompiledHeader="0" \
+                            WarningLevel="3" \
+                            DebugInformationFormat="0" \
+                            Detect64BitPortabilityProblems="true" \

-                $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs"
-                ;;
+                        $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs"
+                    ;;
+                    *)
+                        tag Tool \
+                            Name="VCCLCompilerTool" \
+                            AdditionalIncludeDirectories="$incs" \
+                            PreprocessorDefinitions="WIN32;NDEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
+                            RuntimeLibrary="$release_runtime" \
+                            UsePrecompiledHeader="0" \
+                            WarningLevel="3" \
+                            DebugInformationFormat="0" \
+                            Detect64BitPortabilityProblems="true" \
+
+                        $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs"
+                    ;;
                esac
+            ;;
+        esac

        case "$proj_kind" in
            exe)
                case "$target" in
-                    x86*) tag Tool \
-                                  Name="VCLinkerTool" \
-                                  AdditionalDependencies="$libs \$(NoInherit)" \
-                                  AdditionalLibraryDirectories="$libdirs" \
-                    ;;
-                    arm*|iwmmx*)
+                    x86*)
                        case "$name" in
-                            obj_int_extract) tag Tool \
-                                Name="VCLinkerTool" \
-                                OutputFile="${name}.exe" \
-                                LinkIncremental="1" \
-                                GenerateDebugInformation="false" \
-                                SubSystem="0" \
-                                OptimizeReferences="0" \
-                                EnableCOMDATFolding="0" \
-                                TargetMachine="0"
+                            obj_int_extract)
+                                tag Tool \
+                                    Name="VCLinkerTool" \
+                                    OutputFile="${name}.exe" \
+                                    GenerateDebugInformation="true" \
                            ;;
-                            *) tag Tool \
-                                Name="VCLinkerTool" \
-                                AdditionalDependencies="$libs" \
-                                OutputFile="\$(OutDir)/${name}.exe" \
-                                LinkIncremental="1" \
-                                AdditionalLibraryDirectories="${libdirs};&quot;..\lib/$plat_no_ws&quot;" \
-                                DelayLoadDLLs="\$(NOINHERIT)" \
-                                GenerateDebugInformation="true" \
-                                ProgramDatabaseFile="\$(OutDir)/${name}.pdb" \
-                                SubSystem="9" \
-                                StackReserveSize="65536" \
-                                StackCommitSize="4096" \
-                                OptimizeReferences="2" \
-                                EnableCOMDATFolding="2" \
-                                EntryPointSymbol="mainWCRTStartup" \
-                                TargetMachine="3"
+                            *)
+                                tag Tool \
+                                    Name="VCLinkerTool" \
+                                    AdditionalDependencies="$libs \$(NoInherit)" \
+                                    AdditionalLibraryDirectories="$libdirs" \
+
                            ;;
                        esac
-                     ;;
+                    ;;
                 esac
            ;;
-        lib)
+            lib)
                case "$target" in
-                      arm*|iwmmx*) tag Tool \
-                                    Name="VCLibrarianTool" \
-                                    AdditionalOptions=" /subsystem:windowsce,4.20 /machine:ARM" \
-                                    OutputFile="\$(OutDir)/${name}.lib" \
-                                ;;
-                                *) tag Tool \
-                                    Name="VCLibrarianTool" \
-                                    OutputFile="\$(OutDir)/${name}${lib_sfx}.lib" \
-                                ;;
-                esac
-        ;;
-        dll) # note differences to debug version: LinkIncremental, AssemblyDebug
-             tag Tool \
-                      Name="VCLinkerTool" \
-                      AdditionalDependencies="\$(NoInherit)" \
-                      LinkIncremental="1" \
-                      GenerateDebugInformation="true" \
-                      TargetMachine="1" \
-                      $link_opts
-        esac
+                    x86*)
+                        tag Tool \
+                            Name="VCLibrarianTool" \
+                            OutputFile="\$(OutDir)/${name}${lib_sfx}.lib" \

-        if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
-            case "$name" in
-                vpx)         tag DeploymentTool \
-                             ForceDirty="-1" \
-                             RegisterOutput="0"
-                ;;
-                example|xma) tag DeploymentTool \
-                             ForceDirty="-1" \
-                             RegisterOutput="0"
-                             tag DebuggerTool \
-                             Arguments="${ARGU}"
-                                ;;
-            esac
-        fi
+                    ;;
+                esac
+            ;;
+            dll) # note differences to debug version: LinkIncremental, AssemblyDebug
+                tag Tool \
+                    Name="VCLinkerTool" \
+                    AdditionalDependencies="\$(NoInherit)" \
+                    LinkIncremental="1" \
+                    GenerateDebugInformation="true" \
+                    TargetMachine="1" \
+                    $link_opts \
+
+            ;;
+        esac

        close_tag Configuration
    done
    close_tag Configurations

-    open_tag  Files
-    generate_filter srcs   "Source Files"   "cpp;c;cc;cxx;def;odl;idl;hpj;bat;asm;asmx"
-    generate_filter hdrs   "Header Files"   "h;hpp;hxx;hm;inl;inc;xsd"
+    open_tag Files
+    generate_filter srcs   "Source Files"   "c;def;odl;idl;hpj;bat;asm;asmx"
+    generate_filter hdrs   "Header Files"   "h;hm;inl;inc;xsd"
    generate_filter resrcs "Resource Files" "rc;ico;cur;bmp;dlg;rc2;rct;bin;rgs;gif;jpg;jpeg;jpe;resx;tiff;tif;png;wav"
    generate_filter resrcs "Build Files"    "mk"
    close_tag Files
--- a/build/make/gen_msvs_sln.sh
+++ b/build/make/gen_msvs_sln.sh
@@ -139,9 +139,6 @@ process_global() {
            echo "${indent}${proj_guid}.${config}.ActiveCfg = ${config}"
            echo "${indent}${proj_guid}.${config}.Build.0 = ${config}"

-            if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
-                echo "${indent}${proj_guid}.${config}.Deploy.0 = ${config}"
-            fi
        done
        IFS=${IFS_bak}
    done
--- a/build/make/obj_int_extract.c
+++ b/build/make/obj_int_extract.c
@@ -14,7 +14,7 @@

 #include "vpx_config.h"

-#if defined(_MSC_VER)
+#if defined(_MSC_VER) || defined(__MINGW32__)
 #include <io.h>
 #include <share.h>
 #include "vpx/vpx_integer.h"
@@ -59,20 +59,47 @@ int parse_macho(uint8_t *base_buf, size_t sz)
    struct mach_header header;
    uint8_t *buf = base_buf;
    int base_data_section = 0;
+    int bits = 0;

+    /* We can read in mach_header for 32 and 64 bit architectures
+     * because it's identical to mach_header_64 except for the last
+     * element (uint32_t reserved), which we don't use. Then, when
+     * we know which architecture we're looking at, increment buf
+     * appropriately.
+     */
    memcpy(&header, buf, sizeof(struct mach_header));
-    buf += sizeof(struct mach_header);

-    if (header.magic != MH_MAGIC)
+    if (header.magic == MH_MAGIC)
    {
-        log_msg("Bad magic number for object file. 0x%x expected, 0x%x found.\n",
-                header.magic, MH_MAGIC);
-        goto bail;
+        if (header.cputype == CPU_TYPE_ARM
+            || header.cputype == CPU_TYPE_X86)
+        {
+            bits = 32;
+            buf += sizeof(struct mach_header);
+        }
+        else
+        {
+            log_msg("Bad cputype for object file. Currently only tested for CPU_TYPE_[ARM|X86].\n");
+            goto bail;
+        }
    }
-
-    if (header.cputype != CPU_TYPE_ARM)
+    else if (header.magic == MH_MAGIC_64)
    {
-        log_msg("Bad cputype for object file. Currently only tested for CPU_TYPE_ARM.\n");
+        if (header.cputype == CPU_TYPE_X86_64)
+        {
+            bits = 64;
+            buf += sizeof(struct mach_header_64);
+        }
+        else
+        {
+            log_msg("Bad cputype for object file. Currently only tested for CPU_TYPE_X86_64.\n");
+            goto bail;
+        }
+    }
+    else
+    {
+        log_msg("Bad magic number for object file. 0x%x or 0x%x expected, 0x%x found.\n",
+                MH_MAGIC, MH_MAGIC_64, header.magic);
        goto bail;
    }

@@ -85,8 +112,6 @@ int parse_macho(uint8_t *base_buf, size_t sz)
    for (i = 0; i < header.ncmds; i++)
    {
        struct load_command lc;
-        struct symtab_command sc;
-        struct segment_command seg_c;

        memcpy(&lc, buf, sizeof(struct load_command));

@@ -94,50 +119,99 @@ int parse_macho(uint8_t *base_buf, size_t sz)
        {
            uint8_t *seg_buf = buf;
            struct section s;
+            struct segment_command seg_c;

-            memcpy(&seg_c, buf, sizeof(struct segment_command));
-
+            memcpy(&seg_c, seg_buf, sizeof(struct segment_command));
            seg_buf += sizeof(struct segment_command);

-            for (j = 0; j < seg_c.nsects; j++)
+            /* Although each section is given it's own offset, nlist.n_value
+             * references the offset of the first section. This isn't
+             * apparent without debug information because the offset of the
+             * data section is the same as the first section. However, with
+             * debug sections mixed in, the offset of the debug section
+             * increases but n_value still references the first section.
+             */
+            if (seg_c.nsects < 1)
            {
-                memcpy(&s, seg_buf + (j * sizeof(struct section)), sizeof(struct section));
-
-                // Need to get this offset which is the start of the symbol table
-                // before matching the strings up with symbols.
-                base_data_section = s.offset;
+                log_msg("Not enough sections\n");
+                goto bail;
            }
+
+            memcpy(&s, seg_buf, sizeof(struct section));
+            base_data_section = s.offset;
+        }
+        else if (lc.cmd == LC_SEGMENT_64)
+        {
+            uint8_t *seg_buf = buf;
+            struct section_64 s;
+            struct segment_command_64 seg_c;
+
+            memcpy(&seg_c, seg_buf, sizeof(struct segment_command_64));
+            seg_buf += sizeof(struct segment_command_64);
+
+            /* Explanation in LG_SEGMENT */
+            if (seg_c.nsects < 1)
+            {
+                log_msg("Not enough sections\n");
+                goto bail;
+            }
+
+            memcpy(&s, seg_buf, sizeof(struct section_64));
+            base_data_section = s.offset;
        }
        else if (lc.cmd == LC_SYMTAB)
        {
-            uint8_t *sym_buf = base_buf;
-            uint8_t *str_buf = base_buf;
-
            if (base_data_section != 0)
            {
+                struct symtab_command sc;
+                uint8_t *sym_buf = base_buf;
+                uint8_t *str_buf = base_buf;
+
                memcpy(&sc, buf, sizeof(struct symtab_command));

                if (sc.cmdsize != sizeof(struct symtab_command))
+                {
                    log_msg("Can't find symbol table!\n");
+                    goto bail;
+                }

                sym_buf += sc.symoff;
                str_buf += sc.stroff;

                for (j = 0; j < sc.nsyms; j++)
                {
-                    struct nlist nl;
-                    int val;
+                    /* Location of string is cacluated each time from the
+                     * start of the string buffer.  On darwin the symbols
+                     * are prefixed by "_", so we bump the pointer by 1.
+                     * The target value is defined as an int in asm_*_offsets.c,
+                     * which is 4 bytes on all targets we currently use.
+                     */
+                    if (bits == 32)
+                    {
+                        struct nlist nl;
+                        int val;

-                    memcpy(&nl, sym_buf + (j * sizeof(struct nlist)), sizeof(struct nlist));
+                        memcpy(&nl, sym_buf, sizeof(struct nlist));
+                        sym_buf += sizeof(struct nlist);

-                    val = *((int *)(base_buf + base_data_section + nl.n_value));
+                        memcpy(&val, base_buf + base_data_section + nl.n_value,
+                               sizeof(val));
+                        printf("%-40s EQU %5d\n",
+                               str_buf + nl.n_un.n_strx + 1, val);
+                    }
+                    else /* if (bits == 64) */
+                    {
+                        struct nlist_64 nl;
+                        int val;

-                    // Location of string is cacluated each time from the
-                    // start of the string buffer.  On darwin the symbols
-                    // are prefixed by "_".  On other platforms it is not
-                    // so it needs to be removed.  That is the reason for
-                    // the +1.
-                    printf("%-40s EQU %5d\n", str_buf + nl.n_un.n_strx + 1, val);
+                        memcpy(&nl, sym_buf, sizeof(struct nlist_64));
+                        sym_buf += sizeof(struct nlist_64);
+
+                        memcpy(&val, base_buf + base_data_section + nl.n_value,
+                               sizeof(val));
+                        printf("%-40s EQU %5d\n",
+                               str_buf + nl.n_un.n_strx + 1, val);
+                    }
                }
            }
        }
@@ -218,7 +292,7 @@ bail:
    return EXIT_FAILURE;
 }

-#else
+#elif defined(__ELF__)
 #include "elf.h"

 #define COPY_STRUCT(dst, buf, ofst, sz) do {\
@@ -237,212 +311,420 @@ bail:

 typedef struct
 {
-    uint8_t     *buf; /* Buffer containing ELF data */
-    size_t       sz;  /* Buffer size */
-    int          le_data;   /* Data is little-endian */
-    Elf32_Ehdr   hdr;
+    uint8_t      *buf; /* Buffer containing ELF data */
+    size_t        sz;  /* Buffer size */
+    int           le_data; /* Data is little-endian */
+    unsigned char e_ident[EI_NIDENT]; /* Magic number and other info */
+    int           bits; /* 32 or 64 */
+    Elf32_Ehdr    hdr32;
+    Elf64_Ehdr    hdr64;
 } elf_obj_t;

-int parse_elf32_header(elf_obj_t *elf)
+int parse_elf_header(elf_obj_t *elf)
 {
    int res;
-    /* Verify ELF32 header */
-    COPY_STRUCT(&elf->hdr, elf->buf, 0, elf->sz);
-    res = elf->hdr.e_ident[EI_MAG0] == ELFMAG0;
-    res &= elf->hdr.e_ident[EI_MAG1] == ELFMAG1;
-    res &= elf->hdr.e_ident[EI_MAG2] == ELFMAG2;
-    res &= elf->hdr.e_ident[EI_MAG3] == ELFMAG3;
-    res &= elf->hdr.e_ident[EI_CLASS] == ELFCLASS32;
-    res &= elf->hdr.e_ident[EI_DATA] == ELFDATA2LSB
-           || elf->hdr.e_ident[EI_DATA] == ELFDATA2MSB;
+    /* Verify ELF Magic numbers */
+    COPY_STRUCT(&elf->e_ident, elf->buf, 0, elf->sz);
+    res = elf->e_ident[EI_MAG0] == ELFMAG0;
+    res &= elf->e_ident[EI_MAG1] == ELFMAG1;
+    res &= elf->e_ident[EI_MAG2] == ELFMAG2;
+    res &= elf->e_ident[EI_MAG3] == ELFMAG3;
+    res &= elf->e_ident[EI_CLASS] == ELFCLASS32
+        || elf->e_ident[EI_CLASS] == ELFCLASS64;
+    res &= elf->e_ident[EI_DATA] == ELFDATA2LSB;

    if (!res) goto bail;

-    elf->le_data = elf->hdr.e_ident[EI_DATA] == ELFDATA2LSB;
+    elf->le_data = elf->e_ident[EI_DATA] == ELFDATA2LSB;

-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_type);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_machine);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_version);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_entry);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_phoff);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_shoff);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_flags);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_ehsize);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_phentsize);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_phnum);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_shentsize);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_shnum);
-    ENDIAN_ASSIGN_IN_PLACE(elf->hdr.e_shstrndx);
-    return 0;
-bail:
-    return 1;
-}
-
-int parse_elf32_section(elf_obj_t *elf, int idx, Elf32_Shdr *hdr)
-{
-    if (idx >= elf->hdr.e_shnum)
-        goto bail;
-
-    COPY_STRUCT(hdr, elf->buf, elf->hdr.e_shoff + idx * elf->hdr.e_shentsize,
-                elf->sz);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_name);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_type);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_flags);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_addr);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_offset);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_size);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_link);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_info);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_addralign);
-    ENDIAN_ASSIGN_IN_PLACE(hdr->sh_entsize);
-    return 0;
-bail:
-    return 1;
-}
-
-char *parse_elf32_string_table(elf_obj_t *elf, int s_idx, int idx)
-{
-    Elf32_Shdr shdr;
-
-    if (parse_elf32_section(elf, s_idx, &shdr))
+    /* Read in relevant values */
+    if (elf->e_ident[EI_CLASS] == ELFCLASS32)
    {
-        log_msg("Failed to parse ELF string table: section %d, index %d\n",
-                s_idx, idx);
-        return "";
+        elf->bits = 32;
+        COPY_STRUCT(&elf->hdr32, elf->buf, 0, elf->sz);
+
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_type);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_machine);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_version);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_entry);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_phoff);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_shoff);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_flags);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_ehsize);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_phentsize);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_phnum);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_shentsize);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_shnum);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr32.e_shstrndx);
+    }
+    else /* if (elf->e_ident[EI_CLASS] == ELFCLASS64) */
+    {
+        elf->bits = 64;
+        COPY_STRUCT(&elf->hdr64, elf->buf, 0, elf->sz);
+
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_type);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_machine);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_version);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_entry);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_phoff);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_shoff);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_flags);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_ehsize);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_phentsize);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_phnum);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_shentsize);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_shnum);
+        ENDIAN_ASSIGN_IN_PLACE(elf->hdr64.e_shstrndx);
    }

-    return (char *)(elf->buf + shdr.sh_offset + idx);
+    return 0;
+bail:
+    log_msg("Failed to parse ELF file header");
+    return 1;
 }

-int parse_elf32_symbol(elf_obj_t *elf, unsigned int ofst, Elf32_Sym *sym)
+int parse_elf_section(elf_obj_t *elf, int idx, Elf32_Shdr *hdr32, Elf64_Shdr *hdr64)
 {
-    COPY_STRUCT(sym, elf->buf, ofst, elf->sz);
-    ENDIAN_ASSIGN_IN_PLACE(sym->st_name);
-    ENDIAN_ASSIGN_IN_PLACE(sym->st_value);
-    ENDIAN_ASSIGN_IN_PLACE(sym->st_size);
-    ENDIAN_ASSIGN_IN_PLACE(sym->st_info);
-    ENDIAN_ASSIGN_IN_PLACE(sym->st_other);
-    ENDIAN_ASSIGN_IN_PLACE(sym->st_shndx);
+    if (hdr32)
+    {
+        if (idx >= elf->hdr32.e_shnum)
+            goto bail;
+
+        COPY_STRUCT(hdr32, elf->buf, elf->hdr32.e_shoff + idx * elf->hdr32.e_shentsize,
+                    elf->sz);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_name);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_type);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_flags);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_addr);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_offset);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_size);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_link);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_info);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_addralign);
+        ENDIAN_ASSIGN_IN_PLACE(hdr32->sh_entsize);
+    }
+    else /* if (hdr64) */
+    {
+        if (idx >= elf->hdr64.e_shnum)
+            goto bail;
+
+        COPY_STRUCT(hdr64, elf->buf, elf->hdr64.e_shoff + idx * elf->hdr64.e_shentsize,
+                    elf->sz);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_name);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_type);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_flags);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_addr);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_offset);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_size);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_link);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_info);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_addralign);
+        ENDIAN_ASSIGN_IN_PLACE(hdr64->sh_entsize);
+    }
+
    return 0;
 bail:
    return 1;
 }

-int parse_elf32(uint8_t *buf, size_t sz, output_fmt_t mode)
+char *parse_elf_string_table(elf_obj_t *elf, int s_idx, int idx)
 {
-    elf_obj_t  elf;
-    Elf32_Shdr shdr;
+    if (elf->bits == 32)
+    {
+        Elf32_Shdr shdr;
+
+        if (parse_elf_section(elf, s_idx, &shdr, NULL))
+        {
+            log_msg("Failed to parse ELF string table: section %d, index %d\n",
+                    s_idx, idx);
+            return "";
+        }
+
+        return (char *)(elf->buf + shdr.sh_offset + idx);
+    }
+    else /* if (elf->bits == 64) */
+    {
+        Elf64_Shdr shdr;
+
+        if (parse_elf_section(elf, s_idx, NULL, &shdr))
+        {
+            log_msg("Failed to parse ELF string table: section %d, index %d\n",
+                    s_idx, idx);
+            return "";
+        }
+
+        return (char *)(elf->buf + shdr.sh_offset + idx);
+    }
+}
+
+int parse_elf_symbol(elf_obj_t *elf, unsigned int ofst, Elf32_Sym *sym32, Elf64_Sym *sym64)
+{
+    if (sym32)
+    {
+        COPY_STRUCT(sym32, elf->buf, ofst, elf->sz);
+        ENDIAN_ASSIGN_IN_PLACE(sym32->st_name);
+        ENDIAN_ASSIGN_IN_PLACE(sym32->st_value);
+        ENDIAN_ASSIGN_IN_PLACE(sym32->st_size);
+        ENDIAN_ASSIGN_IN_PLACE(sym32->st_info);
+        ENDIAN_ASSIGN_IN_PLACE(sym32->st_other);
+        ENDIAN_ASSIGN_IN_PLACE(sym32->st_shndx);
+    }
+    else /* if (sym64) */
+    {
+        COPY_STRUCT(sym64, elf->buf, ofst, elf->sz);
+        ENDIAN_ASSIGN_IN_PLACE(sym64->st_name);
+        ENDIAN_ASSIGN_IN_PLACE(sym64->st_value);
+        ENDIAN_ASSIGN_IN_PLACE(sym64->st_size);
+        ENDIAN_ASSIGN_IN_PLACE(sym64->st_info);
+        ENDIAN_ASSIGN_IN_PLACE(sym64->st_other);
+        ENDIAN_ASSIGN_IN_PLACE(sym64->st_shndx);
+    }
+    return 0;
+bail:
+    return 1;
+}
+
+int parse_elf(uint8_t *buf, size_t sz, output_fmt_t mode)
+{
+    elf_obj_t    elf;
    unsigned int ofst;
-    int         i;
-    Elf32_Off strtab_off;   /* save String Table offset for later use */
+    int          i;
+    Elf32_Off    strtab_off32;
+    Elf64_Off    strtab_off64; /* save String Table offset for later use */

    memset(&elf, 0, sizeof(elf));
    elf.buf = buf;
    elf.sz = sz;

    /* Parse Header */
-    if (parse_elf32_header(&elf))
-    {
-        log_msg("Parse error: File does not appear to be valid ELF32\n");
-        return 1;
-    }
+    if (parse_elf_header(&elf))
+      goto bail;

-    for (i = 0; i < elf.hdr.e_shnum; i++)
+    if (elf.bits == 32)
    {
-        parse_elf32_section(&elf, i, &shdr);
-
-        if (shdr.sh_type == SHT_STRTAB)
+        Elf32_Shdr shdr;
+        for (i = 0; i < elf.hdr32.e_shnum; i++)
        {
-            char strtsb_name[128];
+            parse_elf_section(&elf, i, &shdr, NULL);

-            strcpy(strtsb_name, (char *)(elf.buf + shdr.sh_offset + shdr.sh_name));
-
-            if (!(strcmp(strtsb_name, ".shstrtab")))
+            if (shdr.sh_type == SHT_STRTAB)
            {
-                log_msg("found section: %s\n", strtsb_name);
-                strtab_off = shdr.sh_offset;
-                break;
+                char strtsb_name[128];
+
+                strcpy(strtsb_name, (char *)(elf.buf + shdr.sh_offset + shdr.sh_name));
+
+                if (!(strcmp(strtsb_name, ".shstrtab")))
+                {
+                    /* log_msg("found section: %s\n", strtsb_name); */
+                    strtab_off32 = shdr.sh_offset;
+                    break;
+                }
+            }
+        }
+    }
+    else /* if (elf.bits == 64) */
+    {
+        Elf64_Shdr shdr;
+        for (i = 0; i < elf.hdr64.e_shnum; i++)
+        {
+            parse_elf_section(&elf, i, NULL, &shdr);
+
+            if (shdr.sh_type == SHT_STRTAB)
+            {
+                char strtsb_name[128];
+
+                strcpy(strtsb_name, (char *)(elf.buf + shdr.sh_offset + shdr.sh_name));
+
+                if (!(strcmp(strtsb_name, ".shstrtab")))
+                {
+                    /* log_msg("found section: %s\n", strtsb_name); */
+                    strtab_off64 = shdr.sh_offset;
+                    break;
+                }
            }
        }
    }

    /* Parse all Symbol Tables */
-    for (i = 0; i < elf.hdr.e_shnum; i++)
+    if (elf.bits == 32)
    {
-
-        parse_elf32_section(&elf, i, &shdr);
-
-        if (shdr.sh_type == SHT_SYMTAB)
+        Elf32_Shdr shdr;
+        for (i = 0; i < elf.hdr32.e_shnum; i++)
        {
-            for (ofst = shdr.sh_offset;
-                 ofst < shdr.sh_offset + shdr.sh_size;
-                 ofst += shdr.sh_entsize)
+            parse_elf_section(&elf, i, &shdr, NULL);
+
+            if (shdr.sh_type == SHT_SYMTAB)
            {
-                Elf32_Sym sym;
-
-                parse_elf32_symbol(&elf, ofst, &sym);
-
-                /* For all OBJECTS (data objects), extract the value from the
-                 * proper data segment.
-                 */
-                if (ELF32_ST_TYPE(sym.st_info) == STT_OBJECT && sym.st_name)
-                    log_msg("found data object %s\n",
-                            parse_elf32_string_table(&elf,
-                                                     shdr.sh_link,
-                                                     sym.st_name));
-
-                if (ELF32_ST_TYPE(sym.st_info) == STT_OBJECT
-                    && sym.st_size == 4)
+                for (ofst = shdr.sh_offset;
+                     ofst < shdr.sh_offset + shdr.sh_size;
+                     ofst += shdr.sh_entsize)
                {
-                    Elf32_Shdr dhdr;
-                    int32_t      val;
-                    char section_name[128];
+                    Elf32_Sym sym;

-                    parse_elf32_section(&elf, sym.st_shndx, &dhdr);
+                    parse_elf_symbol(&elf, ofst, &sym, NULL);

-                    /* For explanition - refer to _MSC_VER version of code */
-                    strcpy(section_name, (char *)(elf.buf + strtab_off + dhdr.sh_name));
-                    log_msg("Section_name: %s, Section_type: %d\n", section_name, dhdr.sh_type);
+                    /* For all OBJECTS (data objects), extract the value from the
+                     * proper data segment.
+                     */
+                    /* if (ELF32_ST_TYPE(sym.st_info) == STT_OBJECT && sym.st_name)
+                        log_msg("found data object %s\n",
+                                parse_elf_string_table(&elf,
+                                                       shdr.sh_link,
+                                                       sym.st_name));
+                     */

-                    if (!(strcmp(section_name, ".bss")))
+                    if (ELF32_ST_TYPE(sym.st_info) == STT_OBJECT
+                        && sym.st_size == 4)
                    {
-                        val = 0;
-                    }
-                    else
-                    {
-                        memcpy(&val,
-                               elf.buf + dhdr.sh_offset + sym.st_value,
-                               sizeof(val));
+                        Elf32_Shdr dhdr;
+                        int val = 0;
+                        char section_name[128];
+
+                        parse_elf_section(&elf, sym.st_shndx, &dhdr, NULL);
+
+                        /* For explanition - refer to _MSC_VER version of code */
+                        strcpy(section_name, (char *)(elf.buf + strtab_off32 + dhdr.sh_name));
+                        /* log_msg("Section_name: %s, Section_type: %d\n", section_name, dhdr.sh_type); */
+
+                        if (strcmp(section_name, ".bss"))
+                        {
+                            if (sizeof(val) != sym.st_size)
+                            {
+                                /* The target value is declared as an int in
+                                 * asm_*_offsets.c, which is 4 bytes on all
+                                 * targets we currently use. Complain loudly if
+                                 * this is not true.
+                                 */
+                                log_msg("Symbol size is wrong\n");
+                                goto bail;
+                            }
+
+                            memcpy(&val,
+                                   elf.buf + dhdr.sh_offset + sym.st_value,
+                                   sym.st_size);
+                        }
+
+                        if (!elf.le_data)
+                        {
+                            log_msg("Big Endian data not supported yet!\n");
+                            goto bail;
+                        }
+
+                        switch (mode)
+                        {
+                            case OUTPUT_FMT_RVDS:
+                                printf("%-40s EQU %5d\n",
+                                       parse_elf_string_table(&elf,
+                                                              shdr.sh_link,
+                                                              sym.st_name),
+                                       val);
+                                break;
+                            case OUTPUT_FMT_GAS:
+                                printf(".equ %-40s, %5d\n",
+                                       parse_elf_string_table(&elf,
+                                                              shdr.sh_link,
+                                                              sym.st_name),
+                                       val);
+                                break;
+                            default:
+                                printf("%s = %d\n",
+                                       parse_elf_string_table(&elf,
+                                                              shdr.sh_link,
+                                                              sym.st_name),
+                                       val);
+                        }
                    }
+                }
+            }
+        }
+    }
+    else /* if (elf.bits == 64) */
+    {
+        Elf64_Shdr shdr;
+        for (i = 0; i < elf.hdr64.e_shnum; i++)
+        {
+            parse_elf_section(&elf, i, NULL, &shdr);

-                    if (!elf.le_data)
-                    {
-                        log_msg("Big Endian data not supported yet!\n");
-                        goto bail;
-                    }\
+            if (shdr.sh_type == SHT_SYMTAB)
+            {
+                for (ofst = shdr.sh_offset;
+                     ofst < shdr.sh_offset + shdr.sh_size;
+                     ofst += shdr.sh_entsize)
+                {
+                    Elf64_Sym sym;

-                    switch (mode)
+                    parse_elf_symbol(&elf, ofst, NULL, &sym);
+
+                    /* For all OBJECTS (data objects), extract the value from the
+                     * proper data segment.
+                     */
+                    /* if (ELF64_ST_TYPE(sym.st_info) == STT_OBJECT && sym.st_name)
+                        log_msg("found data object %s\n",
+                                parse_elf_string_table(&elf,
+                                                       shdr.sh_link,
+                                                       sym.st_name));
+                     */
+
+                    if (ELF64_ST_TYPE(sym.st_info) == STT_OBJECT
+                        && sym.st_size == 4)
                    {
-                    case OUTPUT_FMT_RVDS:
-                        printf("%-40s EQU %5d\n",
-                               parse_elf32_string_table(&elf,
-                                                        shdr.sh_link,
-                                                        sym.st_name),
-                               val);
-                        break;
-                    case OUTPUT_FMT_GAS:
-                        printf(".equ %-40s, %5d\n",
-                               parse_elf32_string_table(&elf,
-                                                        shdr.sh_link,
-                                                        sym.st_name),
-                               val);
-                        break;
-                    default:
-                        printf("%s = %d\n",
-                               parse_elf32_string_table(&elf,
-                                                        shdr.sh_link,
-                                                        sym.st_name),
-                               val);
+                        Elf64_Shdr dhdr;
+                        int val = 0;
+                        char section_name[128];
+
+                        parse_elf_section(&elf, sym.st_shndx, NULL, &dhdr);
+
+                        /* For explanition - refer to _MSC_VER version of code */
+                        strcpy(section_name, (char *)(elf.buf + strtab_off64 + dhdr.sh_name));
+                        /* log_msg("Section_name: %s, Section_type: %d\n", section_name, dhdr.sh_type); */
+
+                        if ((strcmp(section_name, ".bss")))
+                        {
+                            if (sizeof(val) != sym.st_size)
+                            {
+                                /* The target value is declared as an int in
+                                 * asm_*_offsets.c, which is 4 bytes on all
+                                 * targets we currently use. Complain loudly if
+                                 * this is not true.
+                                 */
+                                log_msg("Symbol size is wrong\n");
+                                goto bail;
+                            }
+
+                            memcpy(&val,
+                                   elf.buf + dhdr.sh_offset + sym.st_value,
+                                   sym.st_size);
+                        }
+
+                        if (!elf.le_data)
+                        {
+                            log_msg("Big Endian data not supported yet!\n");
+                            goto bail;
+                        }
+
+                        switch (mode)
+                        {
+                            case OUTPUT_FMT_RVDS:
+                                printf("%-40s EQU %5d\n",
+                                       parse_elf_string_table(&elf,
+                                                              shdr.sh_link,
+                                                              sym.st_name),
+                                       val);
+                                break;
+                            case OUTPUT_FMT_GAS:
+                                printf(".equ %-40s, %5d\n",
+                                       parse_elf_string_table(&elf,
+                                                              shdr.sh_link,
+                                                              sym.st_name),
+                                       val);
+                                break;
+                            default:
+                                printf("%s = %d\n",
+                                       parse_elf_string_table(&elf,
+                                                              shdr.sh_link,
+                                                              sym.st_name),
+                                       val);
+                        }
                    }
                }
            }
@@ -454,7 +736,7 @@ int parse_elf32(uint8_t *buf, size_t sz, output_fmt_t mode)

    return 0;
 bail:
-    log_msg("Parse error: File does not appear to be valid ELF32\n");
+    log_msg("Parse error: File does not appear to be valid ELF32 or ELF64\n");
    return 1;
 }

@@ -521,8 +803,7 @@ int main(int argc, char **argv)
        goto bail;
    }

-    res = parse_elf32(file_buf, stat_buf.st_size, mode);
-    //res = parse_coff(file_buf, stat_buf.st_size);
+    res = parse_elf(file_buf, stat_buf.st_size, mode);
    free(file_buf);

    if (!res)
@@ -535,7 +816,7 @@ bail:
 #endif


-#if defined(_MSC_VER)
+#if defined(_MSC_VER) || defined(__MINGW32__)
 /*  See "Microsoft Portable Executable and Common Object File Format Specification"
    for reference.
 */
@@ -549,7 +830,6 @@ int parse_coff(unsigned __int8 *buf, size_t sz)
    unsigned int i;
    unsigned __int8 *ptr;
    unsigned __int32 symoffset;
-    FILE *fp;

    char **sectionlist;  //this array holds all section names in their correct order.
    //it is used to check if the symbol is in .bss or .data section.
@@ -560,9 +840,18 @@ int parse_coff(unsigned __int8 *buf, size_t sz)
    strtab_ptr = symtab_ptr + symtab_sz * 18;

    if (nsections > 96)
-        goto bail;
+    {
+        log_msg("Too many sections\n");
+        return 1;
+    }

-    sectionlist = malloc(nsections * sizeof * sectionlist);
+    sectionlist = malloc(nsections * sizeof(sectionlist));
+
+    if (sectionlist == NULL)
+    {
+        log_msg("Allocating first level of section list failed\n");
+        return 1;
+    }

    //log_msg("COFF: Found %u symbols in %u sections.\n", symtab_sz, nsections);

@@ -580,6 +869,12 @@ int parse_coff(unsigned __int8 *buf, size_t sz)
        //log_msg("COFF: Parsing section %s\n",sectionname);

        sectionlist[i] = malloc(strlen(sectionname) + 1);
+
+        if (sectionlist[i] == NULL)
+        {
+            log_msg("Allocating storage for %s failed\n", sectionname);
+            goto bail;
+        }
        strcpy(sectionlist[i], sectionname);

        if (!strcmp(sectionname, ".data")) sectionrawdata_ptr = get_le32(ptr + 20);
@@ -590,14 +885,6 @@ int parse_coff(unsigned __int8 *buf, size_t sz)
    //log_msg("COFF: Symbol table at offset %u\n", symtab_ptr);
    //log_msg("COFF: raw data pointer ofset for section .data is %u\n", sectionrawdata_ptr);

-    fp = fopen("assembly_offsets.asm", "w");
-
-    if (fp == NULL)
-    {
-        perror("open file");
-        goto bail;
-    }
-
    /*  The compiler puts the data with non-zero offset in .data section, but puts the data with
        zero offset in .bss section. So, if the data in in .bss section, set offset=0.
        Note from Wiki: In an object module compiled from C, the bss section contains
@@ -631,13 +918,23 @@ int parse_coff(unsigned __int8 *buf, size_t sz)
                char name[9] = {0, 0, 0, 0, 0, 0, 0, 0, 0};
                strncpy(name, ptr, 8);
                //log_msg("COFF: Parsing symbol %s\n",name);
-                fprintf(fp, "%-40s EQU ", name);
+                /* The 64bit Windows compiler doesn't prefix with an _.
+                 * Check what's there, and bump if necessary
+                 */
+                if (name[0] == '_')
+                    printf("%-40s EQU ", name + 1);
+                else
+                    printf("%-40s EQU ", name);
            }
            else
            {
                //log_msg("COFF: Parsing symbol %s\n",
                //        buf + strtab_ptr + get_le32(ptr+4));
-                fprintf(fp, "%-40s EQU ", buf + strtab_ptr + get_le32(ptr + 4));
+                if ((buf + strtab_ptr + get_le32(ptr + 4))[0] == '_')
+                    printf("%-40s EQU ",
+                           buf + strtab_ptr + get_le32(ptr + 4) + 1);
+                else
+                    printf("%-40s EQU ", buf + strtab_ptr + get_le32(ptr + 4));
            }

            if (!(strcmp(sectionlist[section-1], ".bss")))
@@ -654,14 +951,13 @@ int parse_coff(unsigned __int8 *buf, size_t sz)
            //log_msg("      Address: %u\n",get_le32(ptr+8));
            //log_msg("      Offset: %u\n", symoffset);

-            fprintf(fp, "%5d\n", symoffset);
+            printf("%5d\n", symoffset);
        }

        ptr += 18;
    }

-    fprintf(fp, "    END\n");
-    fclose(fp);
+    printf("    END\n");

    for (i = 0; i < nsections; i++)
    {
@@ -711,11 +1007,7 @@ int main(int argc, char **argv)
    else
        f = argv[1];

-    if (_sopen_s(&fd, f, _O_BINARY, _SH_DENYNO, _S_IREAD | _S_IWRITE))
-    {
-        perror("Unable to open file");
-        goto bail;
-    }
+    fd = _sopen(f, _O_BINARY, _SH_DENYNO, _S_IREAD | _S_IWRITE);

    if (_fstat(fd, &stat_buf))
    {
--- a/build/x86-msvs/obj_int_extract.bat
+++ b/build/x86-msvs/obj_int_extract.bat
@@ -0,0 +1,15 @@
+REM   Copyright (c) 2011 The WebM project authors. All Rights Reserved.
+REM
+REM   Use of this source code is governed by a BSD-style license
+REM   that can be found in the LICENSE file in the root of the source
+REM   tree. An additional intellectual property rights grant can be found
+REM   in the file PATENTS.  All contributing project authors may
+REM   be found in the AUTHORS file in the root of the source tree.
+echo on
+
+cl /I "./" /I "%1" /nologo /c "%1/vp8/common/asm_com_offsets.c"
+cl /I "./" /I "%1" /nologo /c "%1/vp8/decoder/asm_dec_offsets.c"
+cl /I "./" /I "%1" /nologo /c "%1/vp8/encoder/asm_enc_offsets.c"
+obj_int_extract.exe rvds "asm_com_offsets.obj" > "asm_com_offsets.asm"
+obj_int_extract.exe rvds "asm_dec_offsets.obj" > "asm_dec_offsets.asm"
+obj_int_extract.exe rvds "asm_enc_offsets.obj" > "asm_enc_offsets.asm"
--- a/4
+++ b/4
@@ -80,19 +80,15 @@ all_platforms="${all_platforms} armv5te-linux-rvct"
 all_platforms="${all_platforms} armv5te-linux-gcc"
 all_platforms="${all_platforms} armv5te-none-rvct"
 all_platforms="${all_platforms} armv5te-symbian-gcc"
-all_platforms="${all_platforms} armv5te-wince-vs8"
 all_platforms="${all_platforms} armv6-darwin-gcc"
 all_platforms="${all_platforms} armv6-linux-rvct"
 all_platforms="${all_platforms} armv6-linux-gcc"
 all_platforms="${all_platforms} armv6-none-rvct"
 all_platforms="${all_platforms} armv6-symbian-gcc"
-all_platforms="${all_platforms} armv6-wince-vs8"
 all_platforms="${all_platforms} iwmmxt-linux-rvct"
 all_platforms="${all_platforms} iwmmxt-linux-gcc"
-all_platforms="${all_platforms} iwmmxt-wince-vs8"
 all_platforms="${all_platforms} iwmmxt2-linux-rvct"
 all_platforms="${all_platforms} iwmmxt2-linux-gcc"
-all_platforms="${all_platforms} iwmmxt2-wince-vs8"
 all_platforms="${all_platforms} armv7-darwin-gcc"    #neon Cortex-A8
 all_platforms="${all_platforms} armv7-linux-rvct"    #neon Cortex-A8
 all_platforms="${all_platforms} armv7-linux-gcc"     #neon Cortex-A8
--- a/docs.mk
+++ b/docs.mk
@@ -34,7 +34,8 @@ TXT_DOX = $(call enabled,TXT_DOX)

 EXAMPLE_PATH += $(SRC_PATH_BARE) #for CHANGELOG, README, etc

-doxyfile: libs.doxy_template libs.doxy examples.doxy
+doxyfile: $(if $(findstring examples, $(ALL_TARGETS)),examples.doxy)
+doxyfile: libs.doxy_template libs.doxy
 	@echo "    [CREATE] $@"
 	@cat $^ > $@
 	@echo "STRIP_FROM_PATH += $(SRC_PATH_BARE) $(BUILD_ROOT)" >> $@
--- a/examples/decoder_tmpl.c
+++ b/examples/decoder_tmpl.c
@@ -19,7 +19,7 @@
 #define VPX_CODEC_DISABLE_COMPAT 1
 #include "vpx/vpx_decoder.h"
 #include "vpx/vp8dx.h"
-#define interface (&vpx_codec_vp8_dx_algo)
+#define interface (vpx_codec_vp8_dx())
@EXTRA_INCLUDES


--- a/examples/decoder_tmpl.txt
+++ b/examples/decoder_tmpl.txt
@@ -2,7 +2,7 @@
 #define VPX_CODEC_DISABLE_COMPAT 1
 #include "vpx/vpx_decoder.h"
 #include "vpx/vp8dx.h"
-#define interface (&vpx_codec_vp8_dx_algo)
+#define interface (vpx_codec_vp8_dx())
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ DEC_INCLUDES


--- a/examples/encoder_tmpl.c
+++ b/examples/encoder_tmpl.c
@@ -19,7 +19,7 @@
 #define VPX_CODEC_DISABLE_COMPAT 1
 #include "vpx/vpx_encoder.h"
 #include "vpx/vp8cx.h"
-#define interface (&vpx_codec_vp8_cx_algo)
+#define interface (vpx_codec_vp8_cx())
 #define fourcc    0x30385056
@EXTRA_INCLUDES

--- a/examples/encoder_tmpl.txt
+++ b/examples/encoder_tmpl.txt
@@ -2,7 +2,7 @@
 #define VPX_CODEC_DISABLE_COMPAT 1
 #include "vpx/vpx_encoder.h"
 #include "vpx/vp8cx.h"
-#define interface (&vpx_codec_vp8_cx_algo)
+#define interface (vpx_codec_vp8_cx())
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ENC_INCLUDES


--- a/examples/simple_decoder.txt
+++ b/examples/simple_decoder.txt
@@ -33,7 +33,7 @@ Initializing The Codec
 ----------------------
 The decoder is initialized by the following code. This is an example for
 the VP8 decoder, but the code is analogous for all algorithms. Replace
-`&vpx_codec_vp8_dx_algo` with a pointer to the interface exposed by the
+`vpx_codec_vp8_dx()` with a pointer to the interface exposed by the
 algorithm you want to use. The `cfg` argument is left as NULL in this
 example, because we want the algorithm to determine the stream
 configuration (width/height) and allocate memory automatically. This
--- a/libs.mk
+++ b/libs.mk
@@ -9,7 +9,13 @@
 ##


-ASM:=$(if $(filter yes,$(CONFIG_GCC)),.asm.s,.asm)
+# ARM assembly files are written in RVCT-style. We use some make magic to
+# filter those files to allow GCC compilation
+ifeq ($(ARCH_ARM),yes)
+  ASM:=$(if $(filter yes,$(CONFIG_GCC)),.asm.s,.asm)
+else
+  ASM:=.asm
+endif

 CODEC_SRCS-yes += libs.mk

@@ -126,28 +132,22 @@ INSTALL-SRCS-$(CONFIG_CODEC_SRCS) += $(call enabled,CODEC_EXPORTS)
 ifeq ($(CONFIG_EXTERNAL_BUILD),yes)
 ifeq ($(CONFIG_MSVS),yes)

-ifeq ($(ARCH_ARM),yes)
-ifeq ($(HAVE_ARMV5TE),yes)
-ARM_ARCH=v5
-endif
-ifeq ($(HAVE_ARMV6),yes)
-ARM_ARCH=v6
-endif
 obj_int_extract.vcproj: $(SRC_PATH_BARE)/build/make/obj_int_extract.c
-	@cp $(SRC_PATH_BARE)/build/arm-wince-vs8/obj_int_extract.bat .
+	@cp $(SRC_PATH_BARE)/build/x86-msvs/obj_int_extract.bat .
 	@echo "    [CREATE] $@"
-	$(SRC_PATH_BARE)/build/make/gen_msvs_proj.sh\
-			--exe\
-			--target=$(TOOLCHAIN)\
-            $(if $(CONFIG_STATIC_MSVCRT),--static-crt) \
-            --name=obj_int_extract\
-            --proj-guid=E1360C65-D375-4335-8057-7ED99CC3F9B2\
-            --out=$@ $^\
-            -I".&quot;;&quot;$(SRC_PATH_BARE)"
+	$(SRC_PATH_BARE)/build/make/gen_msvs_proj.sh \
+    --exe \
+    --target=$(TOOLCHAIN) \
+    --name=obj_int_extract \
+    --ver=$(CONFIG_VS_VERSION) \
+    --proj-guid=E1360C65-D375-4335-8057-7ED99CC3F9B2 \
+    $(if $(CONFIG_STATIC_MSVCRT),--static-crt) \
+    --out=$@ $^ \
+    -I. \
+    -I"$(SRC_PATH_BARE)" \

 PROJECTS-$(BUILD_LIBVPX) += obj_int_extract.vcproj
 PROJECTS-$(BUILD_LIBVPX) += obj_int_extract.bat
-endif

 vpx.def: $(call enabled,CODEC_EXPORTS)
 	@echo "    [CREATE] $@"
@@ -158,15 +158,16 @@ CLEAN-OBJS += vpx.def

 vpx.vcproj: $(CODEC_SRCS) vpx.def
 	@echo "    [CREATE] $@"
-	$(SRC_PATH_BARE)/build/make/gen_msvs_proj.sh\
-			--lib\
-			--target=$(TOOLCHAIN)\
+	$(SRC_PATH_BARE)/build/make/gen_msvs_proj.sh \
+			--lib \
+			--target=$(TOOLCHAIN) \
            $(if $(CONFIG_STATIC_MSVCRT),--static-crt) \
-            --name=vpx\
-            --proj-guid=DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74\
-            --module-def=vpx.def\
-            --ver=$(CONFIG_VS_VERSION)\
-            --out=$@ $(CFLAGS) $^\
+            --name=vpx \
+            --proj-guid=DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74 \
+            --module-def=vpx.def \
+            --ver=$(CONFIG_VS_VERSION) \
+            --out=$@ $(CFLAGS) $^ \
+            --src-path-bare="$(SRC_PATH_BARE)" \

 PROJECTS-$(BUILD_LIBVPX) += vpx.vcproj

@@ -230,36 +231,38 @@ endif
 #
 # Add assembler dependencies for configuration and offsets
 #
-$(filter %.s.o,$(OBJS-yes)):   $(BUILD_PFX)vpx_config.asm
-$(filter %.asm.o,$(OBJS-yes)): $(BUILD_PFX)vpx_config.asm
+$(filter %.s.o,$(OBJS-yes)):     $(BUILD_PFX)vpx_config.asm
+$(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)vpx_config.asm

 #
 # Calculate platform- and compiler-specific offsets for hand coded assembly
 #
-ifeq ($(ARCH_ARM), yes)
-  asm_com_offsets.asm: obj_int_extract
-  asm_com_offsets.asm: $(VP8_PREFIX)common/asm_com_offsets.c.o
+ifeq ($(CONFIG_EXTERNAL_BUILD),) # Visual Studio uses obj_int_extract.bat
+  ifeq ($(ARCH_ARM), yes)
+    asm_com_offsets.asm: obj_int_extract
+    asm_com_offsets.asm: $(VP8_PREFIX)common/asm_com_offsets.c.o
 	./obj_int_extract rvds $< $(ADS2GAS) > $@
-  OBJS-yes += $(VP8_PREFIX)common/asm_com_offsets.c.o
-  CLEAN-OBJS += asm_com_offsets.asm
-  $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_com_offsets.asm
+    OBJS-yes += $(VP8_PREFIX)common/asm_com_offsets.c.o
+    CLEAN-OBJS += asm_com_offsets.asm
+    $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_com_offsets.asm

-  ifeq ($(CONFIG_VP8_ENCODER), yes)
-    asm_enc_offsets.asm: obj_int_extract
-    asm_enc_offsets.asm: $(VP8_PREFIX)encoder/asm_enc_offsets.c.o
+    ifeq ($(CONFIG_VP8_ENCODER), yes)
+      asm_enc_offsets.asm: obj_int_extract
+      asm_enc_offsets.asm: $(VP8_PREFIX)encoder/asm_enc_offsets.c.o
 	./obj_int_extract rvds $< $(ADS2GAS) > $@
-    OBJS-yes += $(VP8_PREFIX)encoder/asm_enc_offsets.c.o
-    CLEAN-OBJS += asm_enc_offsets.asm
-    $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_enc_offsets.asm
-  endif
+      OBJS-yes += $(VP8_PREFIX)encoder/asm_enc_offsets.c.o
+      CLEAN-OBJS += asm_enc_offsets.asm
+      $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_enc_offsets.asm
+    endif

-  ifeq ($(CONFIG_VP8_DECODER), yes)
-    asm_dec_offsets.asm: obj_int_extract
-    asm_dec_offsets.asm: $(VP8_PREFIX)decoder/asm_dec_offsets.c.o
+    ifeq ($(CONFIG_VP8_DECODER), yes)
+      asm_dec_offsets.asm: obj_int_extract
+      asm_dec_offsets.asm: $(VP8_PREFIX)decoder/asm_dec_offsets.c.o
 	./obj_int_extract rvds $< $(ADS2GAS) > $@
-    OBJS-yes += $(VP8_PREFIX)decoder/asm_dec_offsets.c.o
-    CLEAN-OBJS += asm_dec_offsets.asm
-    $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_dec_offsets.asm
+      OBJS-yes += $(VP8_PREFIX)decoder/asm_dec_offsets.c.o
+      CLEAN-OBJS += asm_dec_offsets.asm
+      $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_dec_offsets.asm
+    endif
  endif
 endif

--- a/mainpage.dox
+++ b/mainpage.dox
@@ -31,7 +31,7 @@
  The WebM project is an open source project supported by its community. For
  questions about this SDK, please mail the apps-devel@webmproject.org list.
  To contribute, see http://www.webmproject.org/code/contribute and mail
-  vpx-devel@webmproject.org.
+  codec-devel@webmproject.org.
 */

 /*!\page changelog CHANGELOG
--- a/solution.mk
+++ b/solution.mk
@@ -9,38 +9,13 @@
 ##


-ifeq ($(ARCH_ARM),yes)
-ARM_DEVELOP=no
-ARM_DEVELOP:=$(if $(filter %vpx.vcproj,$(wildcard *.vcproj)),yes)
-
-ifeq ($(ARM_DEVELOP),yes)
-vpx.sln:
-	@echo "    [COPY] $@"
-	@cp $(SRC_PATH_BARE)/build/arm-wince-vs8/vpx.sln .
-PROJECTS-yes += vpx.sln
-else
-vpx.sln: $(wildcard *.vcproj)
-	@echo "    [CREATE] $@"
-	$(SRC_PATH_BARE)/build/make/gen_msvs_sln.sh \
-            $(if $(filter %vpx.vcproj,$^),--dep=vpxdec:vpx) \
-            $(if $(filter %vpx.vcproj,$^),--dep=xma:vpx) \
-            --ver=$(CONFIG_VS_VERSION)\
-            --target=$(TOOLCHAIN)\
-            --out=$@ $^
-vpx.sln.mk: vpx.sln
-	@true
-
-PROJECTS-yes += vpx.sln vpx.sln.mk
-include vpx.sln.mk
-endif
-
-else
 vpx.sln: $(wildcard *.vcproj)
 	@echo "    [CREATE] $@"
 	$(SRC_PATH_BARE)/build/make/gen_msvs_sln.sh \
            $(if $(filter %vpx.vcproj,$^),\
-                $(foreach vcp,$(filter-out %vpx.vcproj,$^),\
+                $(foreach vcp,$(filter-out %vpx.vcproj %obj_int_extract.vcproj,$^),\
                  --dep=$(vcp:.vcproj=):vpx)) \
+            --dep=vpx:obj_int_extract \
            --ver=$(CONFIG_VS_VERSION)\
            --out=$@ $^
 vpx.sln.mk: vpx.sln
@@ -48,7 +23,6 @@ vpx.sln.mk: vpx.sln

 PROJECTS-yes += vpx.sln vpx.sln.mk
 -include vpx.sln.mk
-endif

 # Always install this file, as it is an unconditional post-build rule.
 INSTALL_MAPS += src/%     $(SRC_PATH_BARE)/%
--- a/usage.dox
+++ b/usage.dox
@@ -25,7 +25,7 @@
    codec may write into to store details about a single instance of that codec.
    Most of the context is implementation specific, and thus opaque to the
    application. The context structure as seen by the application is of fixed
-    size, and thus can be allocated eith with automatic storage or dynamically
+    size, and thus can be allocated with automatic storage or dynamically
    on the heap.

    Most operations require an initialized codec context. Codec context
@@ -74,7 +74,7 @@
    the ABI is versioned. The ABI version number must be passed at
    initialization time to ensure the application is using a header file that
    matches the library. The current ABI version number is stored in the
-    prepropcessor macros #VPX_CODEC_ABI_VERSION, #VPX_ENCODER_ABI_VERSION, and
+    preprocessor macros #VPX_CODEC_ABI_VERSION, #VPX_ENCODER_ABI_VERSION, and
    #VPX_DECODER_ABI_VERSION. For convenience, each initialization function has
    a wrapper macro that inserts the correct version number. These macros are
    named like the initialization methods, but without the _ver suffix.
@@ -125,7 +125,7 @@

    The special value <code>0</code> is reserved to represent an infinite
    deadline. In this case, the codec will perform as much processing as
-    possible to yeild the highest quality frame.
+    possible to yield the highest quality frame.

    By convention, the value <code>1</code> is used to mean "return as fast as
    possible."
@@ -135,7 +135,7 @@

 /*! \page usage_xma External Memory Allocation
    Applications that wish to have fine grained control over how and where
-    decoders allocate memory \ref MAY make use of the e_xternal Memory Allocation
+    decoders allocate memory \ref MAY make use of the eXternal Memory Allocation
    (XMA) interface. Not all codecs support the XMA \ref usage_features.

    To use a decoder in XMA mode, the decoder \ref MUST be initialized with the
@@ -143,7 +143,7 @@
    allocate is heavily dependent on the size of the encoded video frames. The
    size of the video must be known before requesting the decoder's memory map.
    This stream information can be obtained with the vpx_codec_peek_stream_info()
-    function, which does not require a contructed decoder context. If the exact
+    function, which does not require a constructed decoder context. If the exact
    stream is not known, a stream info structure can be created that reflects
    the maximum size that the decoder instance is required to support.

@@ -175,7 +175,7 @@
    \section usage_xma_seg_szalign Segment Size and Alignment
    The sz (size) and align (alignment) parameters describe the required size
    and alignment of the requested segment. Alignment will always be a power of
-    two. Applications \ref MUST honor the aligment requested. Failure to do so
+    two. Applications \ref MUST honor the alignment requested. Failure to do so
    could result in program crashes or may incur a speed penalty.

    \section usage_xma_seg_flags Segment Flags
--- a/vp8/common/alloccommon.c
+++ b/vp8/common/alloccommon.c
@@ -16,12 +16,11 @@
 #include "findnearmv.h"
 #include "entropymode.h"
 #include "systemdependent.h"
-#include "vpxerrors.h"


 extern  void vp8_init_scan_order_mask();

-void vp8_update_mode_info_border(MODE_INFO *mi, int rows, int cols)
+static void update_mode_info_border(MODE_INFO *mi, int rows, int cols)
 {
    int i;
    vpx_memset(mi - cols - 2, 0, sizeof(MODE_INFO) * (cols + 1));
@@ -71,7 +70,7 @@ int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
      if (vp8_yv12_alloc_frame_buffer(&oci->yv12_fb[i],  width, height, VP8BORDERINPIXELS) < 0)
        {
            vp8_de_alloc_frame_buffers(oci);
-            return ALLOC_FAILURE;
+            return 1;
        }
    }

@@ -88,13 +87,13 @@ int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
    if (vp8_yv12_alloc_frame_buffer(&oci->temp_scale_frame,   width, 16, VP8BORDERINPIXELS) < 0)
    {
        vp8_de_alloc_frame_buffers(oci);
-        return ALLOC_FAILURE;
+        return 1;
    }

    if (vp8_yv12_alloc_frame_buffer(&oci->post_proc_buffer, width, height, VP8BORDERINPIXELS) < 0)
    {
        vp8_de_alloc_frame_buffers(oci);
-        return ALLOC_FAILURE;
+        return 1;
    }

    oci->mb_rows = height >> 4;
@@ -106,7 +105,7 @@ int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
    if (!oci->mip)
    {
        vp8_de_alloc_frame_buffers(oci);
-        return ALLOC_FAILURE;
+        return 1;
    }

    oci->mi = oci->mip + oci->mode_info_stride + 1;
@@ -117,10 +116,10 @@ int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
    if (!oci->above_context)
    {
        vp8_de_alloc_frame_buffers(oci);
-        return ALLOC_FAILURE;
+        return 1;
    }

-    vp8_update_mode_info_border(oci->mi, oci->mb_rows, oci->mb_cols);
+    update_mode_info_border(oci->mi, oci->mb_rows, oci->mb_cols);

    return 0;
 }
--- a/vp8/common/arm/arm_systemdependent.c
+++ b/vp8/common/arm/arm_systemdependent.c
@@ -11,21 +11,13 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/arm.h"
-#include "g_common.h"
-#include "pragmas.h"
-#include "subpixel.h"
-#include "loopfilter.h"
-#include "recon.h"
-#include "idct.h"
-#include "onyxc_int.h"
-
-extern void (*vp8_build_intra_predictors_mby_ptr)(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby_neon(MACROBLOCKD *x);
-
-extern void (*vp8_build_intra_predictors_mby_s_ptr)(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby_s_neon(MACROBLOCKD *x);
+#include "vp8/common/g_common.h"
+#include "vp8/common/pragmas.h"
+#include "vp8/common/subpixel.h"
+#include "vp8/common/loopfilter.h"
+#include "vp8/common/recon.h"
+#include "vp8/common/idct.h"
+#include "vp8/common/onyxc_int.h"

 void vp8_arch_arm_common_init(VP8_COMMON *ctx)
 {
@@ -106,31 +98,12 @@ void vp8_arch_arm_common_init(VP8_COMMON *ctx)
        rtcd->recon.recon2      = vp8_recon2b_neon;
        rtcd->recon.recon4      = vp8_recon4b_neon;
        rtcd->recon.recon_mb    = vp8_recon_mb_neon;
-
+        rtcd->recon.build_intra_predictors_mby =
+            vp8_build_intra_predictors_mby_neon;
+        rtcd->recon.build_intra_predictors_mby_s =
+            vp8_build_intra_predictors_mby_s_neon;
    }
 #endif

 #endif
-
-#if HAVE_ARMV6
-#if CONFIG_RUNTIME_CPU_DETECT
-    if (has_media)
-#endif
-    {
-        vp8_build_intra_predictors_mby_ptr = vp8_build_intra_predictors_mby;
-        vp8_build_intra_predictors_mby_s_ptr = vp8_build_intra_predictors_mby_s;
-    }
-#endif
-
-#if HAVE_ARMV7
-#if CONFIG_RUNTIME_CPU_DETECT
-    if (has_neon)
-#endif
-    {
-        vp8_build_intra_predictors_mby_ptr =
-         vp8_build_intra_predictors_mby_neon;
-        vp8_build_intra_predictors_mby_s_ptr =
-         vp8_build_intra_predictors_mby_s_neon;
-    }
-#endif
 }
--- a/vp8/common/arm/bilinearfilter_arm.c
+++ b/vp8/common/arm/bilinearfilter_arm.c
@@ -10,9 +10,9 @@


 #include <math.h>
-#include "filter.h"
-#include "subpixel.h"
-#include "arm/bilinearfilter_arm.h"
+#include "vp8/common/filter.h"
+#include "vp8/common/subpixel.h"
+#include "bilinearfilter_arm.h"

 void vp8_filter_block2d_bil_armv6
 (
--- a/vp8/common/arm/filter_arm.c
+++ b/vp8/common/arm/filter_arm.c
@@ -11,8 +11,8 @@

 #include "vpx_ports/config.h"
 #include <math.h>
-#include "filter.h"
-#include "subpixel.h"
+#include "vp8/common/filter.h"
+#include "vp8/common/subpixel.h"
 #include "vpx_ports/mem.h"

 extern void vp8_filter_block2d_first_pass_armv6
--- a/vp8/common/arm/loopfilter_arm.c
+++ b/vp8/common/arm/loopfilter_arm.c
@@ -11,8 +11,8 @@

 #include "vpx_ports/config.h"
 #include <math.h>
-#include "loopfilter.h"
-#include "onyxc_int.h"
+#include "vp8/common/loopfilter.h"
+#include "vp8/common/onyxc_int.h"

 extern prototype_loopfilter(vp8_loop_filter_horizontal_edge_armv6);
 extern prototype_loopfilter(vp8_loop_filter_vertical_edge_armv6);
--- a/vp8/common/arm/neon/recon_neon.c
+++ b/vp8/common/arm/neon/recon_neon.c
@@ -10,8 +10,8 @@


 #include "vpx_ports/config.h"
-#include "recon.h"
-#include "blockd.h"
+#include "vp8/common/recon.h"
+#include "vp8/common/blockd.h"

 extern void vp8_recon16x16mb_neon(unsigned char *pred_ptr, short *diff_ptr, unsigned char *dst_ptr, int ystride, unsigned char *udst_ptr, unsigned char *vdst_ptr);

--- a/vp8/common/arm/recon_arm.h
+++ b/vp8/common/arm/recon_arm.h
@@ -53,6 +53,9 @@ extern prototype_copy_block(vp8_copy_mem16x16_neon);

 extern prototype_recon_macroblock(vp8_recon_mb_neon);

+extern prototype_build_intra_predictors(vp8_build_intra_predictors_mby_neon);
+extern prototype_build_intra_predictors(vp8_build_intra_predictors_mby_s_neon);
+
 #if !CONFIG_RUNTIME_CPU_DETECT
 #undef  vp8_recon_recon
 #define vp8_recon_recon vp8_recon_b_neon
@@ -74,6 +77,13 @@ extern prototype_recon_macroblock(vp8_recon_mb_neon);

 #undef  vp8_recon_recon_mb
 #define vp8_recon_recon_mb vp8_recon_mb_neon
+
+#undef  vp8_recon_build_intra_predictors_mby
+#define vp8_recon_build_intra_predictors_mby vp8_build_intra_predictors_mby_neon
+
+#undef  vp8_recon_build_intra_predictors_mby_s
+#define vp8_recon_build_intra_predictors_mby_s vp8_build_intra_predictors_mby_s_neon
+
 #endif
 #endif

--- a/vp8/common/arm/reconintra_arm.c
+++ b/vp8/common/arm/reconintra_arm.c
@@ -10,10 +10,10 @@


 #include "vpx_ports/config.h"
-#include "blockd.h"
-#include "reconintra.h"
+#include "vp8/common/blockd.h"
+#include "vp8/common/reconintra.h"
 #include "vpx_mem/vpx_mem.h"
-#include "recon.h"
+#include "vp8/common/recon.h"

 #if HAVE_ARMV7
 extern void vp8_build_intra_predictors_mby_neon_func(
--- a/vp8/common/blockd.c
+++ b/vp8/common/blockd.c
@@ -12,8 +12,6 @@
 #include "blockd.h"
 #include "vpx_mem/vpx_mem.h"

-const int vp8_block2type[25] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 1};
-
 const unsigned char vp8_block2left[25] =
 {
    0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8
--- a/vp8/common/blockd.h
+++ b/vp8/common/blockd.h
@@ -28,11 +28,6 @@ void vpx_log(const char *format, ...);
 #define DCPREDSIMTHRESH 0
 #define DCPREDCNTTHRESH 3

-#define Y1CONTEXT 0
-#define UCONTEXT 1
-#define VCONTEXT 2
-#define Y2CONTEXT 3
-
 #define MB_FEATURE_TREE_PROBS   3
 #define MAX_MB_SEGMENTS         4

@@ -48,6 +43,11 @@ typedef struct
    int r, c;
 } POS;

+#define PLANE_TYPE_Y_NO_DC    0
+#define PLANE_TYPE_Y2         1
+#define PLANE_TYPE_UV         2
+#define PLANE_TYPE_Y_WITH_DC  3
+

 typedef char ENTROPY_CONTEXT;
 typedef struct
@@ -58,8 +58,6 @@ typedef struct
    ENTROPY_CONTEXT y2;
 } ENTROPY_CONTEXT_PLANES;

-extern const int vp8_block2type[25];
-
 extern const unsigned char vp8_block2left[25];
 extern const unsigned char vp8_block2above[25];

--- a/vp8/common/boolcoder.h
+++ b/vp8/common/boolcoder.h
@@ -1,570 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef bool_coder_h
-#define bool_coder_h 1
-
-/* Arithmetic bool coder with largish probability range.
-   Timothy S Murphy  6 August 2004 */
-
-/* So as not to force users to drag in too much of my idiosyncratic C++ world,
-   I avoid fancy storage management. */
-
-#include <assert.h>
-
-#include <stddef.h>
-#include <stdio.h>
-
-typedef unsigned char vp8bc_index_t; // probability index
-
-/* There are a couple of slight variants in the details of finite-precision
-   arithmetic coding.  May be safely ignored by most users. */
-
-enum vp8bc_rounding
-{
-    vp8bc_down = 0,     // just like VP8
-    vp8bc_down_full = 1, // handles minimum probability correctly
-    vp8bc_up = 2
-};
-
-#if _MSC_VER
-
-/* Note that msvc by default does not inline _anything_ (regardless of the
-   setting of inline_depth) and that a command-line option (-Ob1 or -Ob2)
-   is required to inline even the smallest functions. */
-
-#   pragma inline_depth( 255)           // I mean it when I inline something
-#   pragma warning( disable : 4099)     // No class vs. struct harassment
-#   pragma warning( disable : 4250)     // dominance complaints
-#   pragma warning( disable : 4284)     // operator-> in templates
-#   pragma warning( disable : 4800)     // bool conversion
-
-// don't let prefix ++,-- stand in for postfix, disaster would ensue
-
-#   pragma warning( error : 4620 4621)
-
-#endif  // _MSC_VER
-
-
-#if __cplusplus
-
-// Sometimes one wishes to be definite about integer lengths.
-
-struct int_types
-{
-    typedef const bool cbool;
-    typedef const signed char cchar;
-    typedef const short cshort;
-    typedef const int cint;
-    typedef const int clong;
-
-    typedef const double cdouble;
-    typedef const size_t csize_t;
-
-    typedef unsigned char uchar;    // 8 bits
-    typedef const uchar cuchar;
-
-    typedef short int16;
-    typedef unsigned short uint16;
-    typedef const int16 cint16;
-    typedef const uint16 cuint16;
-
-    typedef int int32;
-    typedef unsigned int uint32;
-    typedef const int32 cint32;
-    typedef const uint32 cuint32;
-
-    typedef unsigned int uint;
-    typedef unsigned int ulong;
-    typedef const uint cuint;
-    typedef const ulong culong;
-
-
-    // All structs consume space, may as well have a vptr.
-
-    virtual ~int_types();
-};
-
-
-struct bool_coder_spec;
-struct bool_coder;
-struct bool_writer;
-struct bool_reader;
-
-
-struct bool_coder_namespace : int_types
-{
-    typedef vp8bc_index_t Index;
-    typedef bool_coder_spec Spec;
-    typedef const Spec c_spec;
-
-    enum Rounding
-    {
-        Down = vp8bc_down,
-        down_full = vp8bc_down_full,
-        Up = vp8bc_up
-    };
-};
-
-
-// Archivable specification of a bool coder includes rounding spec
-// and probability mapping table.  The latter replaces a uchar j
-// (0 <= j < 256) with an arbitrary uint16 tbl[j] = p.
-// p/65536 is then the probability of a zero.
-
-struct bool_coder_spec : bool_coder_namespace
-{
-    friend struct bool_coder;
-    friend struct bool_writer;
-    friend struct bool_reader;
-    friend struct bool_coder_spec_float;
-    friend struct bool_coder_spec_explicit_table;
-    friend struct bool_coder_spec_exponential_table;
-    friend struct BPsrc;
-private:
-    uint w;                 // precision
-    Rounding r;
-
-    uint ebits, mbits, ebias;
-    uint32 mmask;
-
-    Index max_index, half_index;
-
-    uint32 mantissa(Index i) const
-    {
-        assert(i < half_index);
-        return (1 << mbits) + (i & mmask);
-    }
-    uint exponent(Index i) const
-    {
-        assert(i < half_index);
-        return ebias - (i >> mbits);
-    }
-
-    uint16 Ptbl[256];       // kinda clunky, but so is storage management.
-
-    /* Cost in bits of encoding a zero at every probability, scaled by 2^20.
-       Assumes that index is at most 8 bits wide. */
-
-    uint32 Ctbl[256];
-
-    uint32 split(Index i, uint32 R) const    // 1 <= split <= max( 1, R-1)
-    {
-        if (!ebias)
-            return 1 + (((R - 1) * Ptbl[i]) >> 16);
-
-        if (i >= half_index)
-            return R - split(max_index - i, R);
-
-        return 1 + (((R - 1) * mantissa(i)) >> exponent(i));
-    }
-
-    uint32 max_range() const
-    {
-        return (1 << w) - (r == down_full ? 0 : 1);
-    }
-    uint32 min_range() const
-    {
-        return (1 << (w - 1)) + (r == down_full ? 1 : 0);
-    }
-    uint32 Rinc() const
-    {
-        return r == Up ? 1 : 0;
-    }
-
-    void check_prec() const;
-
-    bool float_init(uint Ebits, uint Mbits);
-
-    void cost_init();
-
-    bool_coder_spec(
-        uint prec, Rounding rr, uint Ebits = 0, uint Mbits = 0
-    )
-        : w(prec), r(rr)
-    {
-        float_init(Ebits, Mbits);
-    }
-public:
-    // Read complete spec from file.
-    bool_coder_spec(FILE *);
-
-    // Write spec to file.
-    void dump(FILE *) const;
-
-    // return probability index best approximating prob.
-    Index operator()(double prob) const;
-
-    // probability corresponding to index
-    double operator()(Index i) const;
-
-    Index complement(Index i) const
-    {
-        return max_index - i;
-    }
-
-    Index max_index() const
-    {
-        return max_index;
-    }
-    Index half_index() const
-    {
-        return half_index;
-    }
-
-    uint32 cost_zero(Index i) const
-    {
-        return Ctbl[i];
-    }
-    uint32 cost_one(Index i) const
-    {
-        return Ctbl[ max_index - i];
-    }
-    uint32 cost_bit(Index i, bool b) const
-    {
-        return Ctbl[b? max_index-i:i];
-    }
-};
-
-
-/* Pseudo floating-point probability specification.
-
-   At least one of Ebits and Mbits must be nonzero.
-
-   Since all arithmetic is done at 32 bits, Ebits is at most 5.
-
-   Total significant bits in index is Ebits + Mbits + 1.
-
-   Below the halfway point (i.e. when the top significant bit is 0),
-   the index is (e << Mbits) + m.
-
-   The exponent e is between 0 and (2**Ebits) - 1,
-   the mantissa m is between 0 and (2**Mbits) - 1.
-
-   Prepending an implicit 1 to the mantissa, the probability is then
-
-        (2**Mbits + m) >> (e - 2**Ebits - 1 - Mbits),
-
-   which has (1/2)**(2**Ebits + 1) as a minimum
-   and (1/2) * [1 - 2**(Mbits + 1)] as a maximum.
-
-   When the index is above the halfway point, the probability is the
-   complement of the probability associated to the complement of the index.
-
-   Note that the probability increases with the index and that, because of
-   the symmetry, we cannot encode probability exactly 1/2; though we
-   can get as close to 1/2 as we like, provided we have enough Mbits.
-
-   The latter is of course not a problem in practice, one never has
-   exact probabilities and entropy errors are second order, that is, the
-   "overcoding" of a zero will be largely compensated for by the
-   "undercoding" of a one (or vice-versa).
-
-   Compared to arithmetic probability specs (a la VP8), this will do better
-   at very high and low probabilities and worse at probabilities near 1/2,
-   as well as facilitating the usage of wider or narrower probability indices.
-*/
-
-struct bool_coder_spec_float : bool_coder_spec
-{
-    bool_coder_spec_float(
-        uint Ebits = 3, uint Mbits = 4, Rounding rr = down_full, uint prec = 12
-    )
-        : bool_coder_spec(prec, rr, Ebits, Mbits)
-    {
-        cost_init();
-    }
-};
-
-
-struct bool_coder_spec_explicit_table : bool_coder_spec
-{
-    bool_coder_spec_explicit_table(
-        cuint16 probability_table[256] = 0,  // default is tbl[i] = i << 8.
-        Rounding = down_full,
-        uint precision = 16
-    );
-};
-
-// Contruct table via multiplicative interpolation between
-// p[128] = 1/2  and p[0] = (1/2)^x.
-// Since we are working with 16-bit precision, x is at most 16.
-// For probabilities to increase with i, we must have x > 1.
-// For 0 <= i <= 128, p[i] = (1/2)^{ 1 + [1 - (i/128)]*[x-1] }.
-// Finally, p[128+i] = 1 - p[128 - i].
-
-struct bool_coder_spec_exponential_table : bool_coder_spec
-{
-    bool_coder_spec_exponential_table(uint x, Rounding = down_full, uint prec = 16);
-};
-
-
-// Commonalities between writer and reader.
-
-struct bool_coder : bool_coder_namespace
-{
-    friend struct bool_writer;
-    friend struct bool_reader;
-    friend struct BPsrc;
-private:
-    uint32 Low, Range;
-    cuint32 min_range;
-    cuint32 rinc;
-    c_spec spec;
-
-    void _reset()
-    {
-        Low = 0;
-        Range = spec.max_range();
-    }
-
-    bool_coder(c_spec &s)
-        :  min_range(s.min_range()),
-           rinc(s.Rinc()),
-           spec(s)
-    {
-        _reset();
-    }
-
-    uint32 half() const
-    {
-        return 1 + ((Range - 1) >> 1);
-    }
-public:
-    c_spec &Spec() const
-    {
-        return spec;
-    }
-};
-
-
-struct bool_writer : bool_coder
-{
-    friend struct BPsrc;
-private:
-    uchar *Bstart, *Bend, *B;
-    int bit_lag;
-    bool is_toast;
-    void carry();
-    void reset()
-    {
-        _reset();
-        bit_lag = 32 - spec.w;
-        is_toast = 0;
-    }
-    void raw(bool value, uint32 split);
-public:
-    bool_writer(c_spec &, uchar *Dest, size_t Len);
-    virtual ~bool_writer();
-
-    void operator()(Index p, bool v)
-    {
-        raw(v, spec.split(p, Range));
-    }
-
-    uchar *buf() const
-    {
-        return Bstart;
-    }
-    size_t bytes_written() const
-    {
-        return B - Bstart;
-    }
-
-    // Call when done with input, flushes internal state.
-    // DO NOT write any more data after calling this.
-
-    bool_writer &flush();
-
-    void write_bits(int n, uint val)
-    {
-        if (n)
-        {
-            uint m = 1 << (n - 1);
-
-            do
-            {
-                raw((bool)(val & m), half());
-            }
-            while (m >>= 1);
-        }
-    }
-
-#   if 0
-    // We are agnostic about storage management.
-    // By default, overflows throw an assert but user can
-    // override to provide an expanding buffer using ...
-
-    virtual void overflow(uint Len) const;
-
-    // ... this function copies already-written data into new buffer
-    // and retains new buffer location.
-
-    void new_buffer(uchar *dest, uint Len);
-
-    // Note that storage management is the user's responsibility.
-#   endif
-};
-
-
-// This could be adjusted to use a little less lookahead.
-
-struct bool_reader : bool_coder
-{
-    friend struct BPsrc;
-private:
-    cuchar *const Bstart;   // for debugging
-    cuchar *B;
-    cuchar *const Bend;
-    cuint shf;
-    uint bct;
-    bool raw(uint32 split);
-public:
-    bool_reader(c_spec &s, cuchar *src, size_t Len);
-
-    bool operator()(Index p)
-    {
-        return raw(spec.split(p, Range));
-    }
-
-    uint read_bits(int num_bits)
-    {
-        uint v = 0;
-
-        while (--num_bits >= 0)
-            v += v + (raw(half()) ? 1 : 0);
-
-        return v;
-    }
-};
-
-extern "C" {
-
-#endif /*  __cplusplus */
-
-
-    /* C interface */
-
-    typedef struct bool_coder_spec bool_coder_spec;
-    typedef struct bool_writer bool_writer;
-    typedef struct bool_reader bool_reader;
-
-    typedef const bool_coder_spec c_bool_coder_spec;
-    typedef const bool_writer c_bool_writer;
-    typedef const bool_reader c_bool_reader;
-
-
-    /* Optionally override default precision when constructing coder_specs.
-       Just pass a zero pointer if you don't care.
-       Precision is at most 16 bits for table specs, at most 23 otherwise. */
-
-    struct vp8bc_prec
-    {
-        enum vp8bc_rounding r;      /* see top header file for def */
-        unsigned int prec;          /* range precision in bits */
-    };
-
-    typedef const struct vp8bc_prec vp8bc_c_prec;
-
-    /* bool_coder_spec contains mapping of uchars to actual probabilities
-       (16 bit uints) as well as (usually immaterial) selection of
-       exact finite-precision algorithm used (for now, the latter can only
-       be overridden using the C++ interface).
-       See comments above the corresponding C++ constructors for discussion,
-       especially of exponential probability table generation. */
-
-    bool_coder_spec *vp8bc_vp8spec(); // just like vp8
-
-    bool_coder_spec *vp8bc_literal_spec(
-        const unsigned short prob_map[256],  // 0 is like vp8 w/more precision
-        vp8bc_c_prec*
-    );
-
-    bool_coder_spec *vp8bc_float_spec(
-        unsigned int exponent_bits, unsigned int mantissa_bits, vp8bc_c_prec*
-    );
-
-    bool_coder_spec *vp8bc_exponential_spec(unsigned int min_exp, vp8bc_c_prec *);
-
-    bool_coder_spec *vp8bc_spec_from_file(FILE *);
-
-
-    void vp8bc_destroy_spec(c_bool_coder_spec *);
-
-    void vp8bc_spec_to_file(c_bool_coder_spec *, FILE *);
-
-
-    /* Nearest index to supplied probability of zero, 0 <= prob <= 1. */
-
-    vp8bc_index_t vp8bc_index(c_bool_coder_spec *, double prob);
-
-    vp8bc_index_t vp8bc_index_from_counts(
-        c_bool_coder_spec *p, unsigned int zero_ct, unsigned int one_ct
-    );
-
-    /* In case you want to look */
-
-    double vp8bc_probability(c_bool_coder_spec *, vp8bc_index_t);
-
-    /* Opposite index */
-
-    vp8bc_index_t vp8bc_complement(c_bool_coder_spec *, vp8bc_index_t);
-
-    /* Cost in bits of encoding a zero at given probability, scaled by 2^20.
-       (assumes that an int holds at least 32 bits). */
-
-    unsigned int vp8bc_cost_zero(c_bool_coder_spec *, vp8bc_index_t);
-
-    unsigned int vp8bc_cost_one(c_bool_coder_spec *, vp8bc_index_t);
-    unsigned int vp8bc_cost_bit(c_bool_coder_spec *, vp8bc_index_t, int);
-
-
-    /* bool_writer interface */
-
-    /* Length = 0 disables checking for writes beyond buffer end. */
-
-    bool_writer *vp8bc_create_writer(
-        c_bool_coder_spec *, unsigned char *Destination, size_t Length
-    );
-
-    /* Flushes out any buffered data and returns total # of bytes written. */
-
-    size_t vp8bc_destroy_writer(bool_writer *);
-
-    void vp8bc_write_bool(bool_writer *, int boolean_val, vp8bc_index_t false_prob);
-
-    void vp8bc_write_bits(
-        bool_writer *, unsigned int integer_value, int number_of_bits
-    );
-
-    c_bool_coder_spec *vp8bc_writer_spec(c_bool_writer *);
-
-
-    /* bool_reader interface */
-
-    /* Length = 0 disables checking for reads beyond buffer end. */
-
-    bool_reader *vp8bc_create_reader(
-        c_bool_coder_spec *, const unsigned char *Source, size_t Length
-    );
-    void vp8bc_destroy_reader(bool_reader *);
-
-    int vp8bc_read_bool(bool_reader *, vp8bc_index_t false_prob);
-
-    unsigned int vp8bc_read_bits(bool_reader *, int number_of_bits);
-
-    c_bool_coder_spec *vp8bc_reader_spec(c_bool_reader *);
-
-#if __cplusplus
-}
-#endif
-
-#endif  /* bool_coder_h */
--- a/vp8/common/codec_common_interface.h
+++ b/vp8/common/codec_common_interface.h
@@ -1,93 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-#ifndef CODEC_COMMON_INTERFACE_H
-#define CODEC_COMMON_INTERFACE_H
-
-#define __export
-#define _export
-#define dll_export   __declspec( dllexport )
-#define dll_import   __declspec( dllimport )
-
-// Playback ERROR Codes.
-#define NO_DECODER_ERROR            0
-#define REMOTE_DECODER_ERROR        -1
-
-#define DFR_BAD_DCT_COEFF           -100
-#define DFR_ZERO_LENGTH_FRAME       -101
-#define DFR_FRAME_SIZE_INVALID      -102
-#define DFR_OUTPUT_BUFFER_OVERFLOW  -103
-#define DFR_INVALID_FRAME_HEADER    -104
-#define FR_INVALID_MODE_TOKEN       -110
-#define ETR_ALLOCATION_ERROR        -200
-#define ETR_INVALID_ROOT_PTR        -201
-#define SYNCH_ERROR                 -400
-#define BUFFER_UNDERFLOW_ERROR      -500
-#define PB_IB_OVERFLOW_ERROR        -501
-
-// External error triggers
-#define PB_HEADER_CHECKSUM_ERROR    -601
-#define PB_DATA_CHECKSUM_ERROR      -602
-
-// DCT Error Codes
-#define DDCT_EXPANSION_ERROR        -700
-#define DDCT_INVALID_TOKEN_ERROR    -701
-
-// exception_errors
-#define GEN_EXCEPTIONS              -800
-#define EX_UNQUAL_ERROR             -801
-
-// Unrecoverable error codes
-#define FATAL_PLAYBACK_ERROR        -1000
-#define GEN_ERROR_CREATING_CDC      -1001
-#define GEN_THREAD_CREATION_ERROR   -1002
-#define DFR_CREATE_BMP_FAILED       -1003
-
-// YUV buffer configuration structure
-typedef struct
-{
-    int     y_width;
-    int     y_height;
-    int     y_stride;
-
-    int     uv_width;
-    int     uv_height;
-    int     uv_stride;
-
-    unsigned char   *y_buffer;
-    unsigned char   *u_buffer;
-    unsigned char   *v_buffer;
-
-} YUV_BUFFER_CONFIG;
-typedef enum
-{
-    C_SET_KEY_FRAME,
-    C_SET_FIXED_Q,
-    C_SET_FIRSTPASS_FILE,
-    C_SET_EXPERIMENTAL_MIN,
-    C_SET_EXPERIMENTAL_MAX = C_SET_EXPERIMENTAL_MIN + 255,
-    C_SET_CHECKPROTECT,
-    C_SET_TESTMODE,
-    C_SET_INTERNAL_SIZE,
-    C_SET_RECOVERY_FRAME,
-    C_SET_REFERENCEFRAME,
-    C_SET_GOLDENFRAME
-
-#ifndef VP50_COMP_INTERFACE
-    // Specialist test facilities.
-//    C_VCAP_PARAMS,              // DO NOT USE FOR NOW WITH VFW CODEC
-#endif
-
-} C_SETTING;
-
-typedef unsigned long C_SET_VALUE;
-
-
-#endif
--- a/vp8/common/filter.c
+++ b/vp8/common/filter.c
@@ -38,7 +38,7 @@ DECLARE_ALIGNED(16, const short, vp8_sub_pel_filters[8][6]) =
    { 0, -1,   12,  123,  -6,  0 },
 };

-void vp8_filter_block2d_first_pass
+static void filter_block2d_first_pass
 (
    unsigned char *src_ptr,
    int *output_ptr,
@@ -82,7 +82,7 @@ void vp8_filter_block2d_first_pass
    }
 }

-void vp8_filter_block2d_second_pass
+static void filter_block2d_second_pass
 (
    int *src_ptr,
    unsigned char *output_ptr,
@@ -129,7 +129,7 @@ void vp8_filter_block2d_second_pass
 }


-void vp8_filter_block2d
+static void filter_block2d
 (
    unsigned char  *src_ptr,
    unsigned char  *output_ptr,
@@ -142,39 +142,13 @@ void vp8_filter_block2d
    int FData[9*4]; /* Temp data buffer used in filtering */

    /* First filter 1-D horizontally... */
-    vp8_filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 9, 4, HFilter);
+    filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 9, 4, HFilter);

    /* then filter verticaly... */
-    vp8_filter_block2d_second_pass(FData + 8, output_ptr, output_pitch, 4, 4, 4, 4, VFilter);
+    filter_block2d_second_pass(FData + 8, output_ptr, output_pitch, 4, 4, 4, 4, VFilter);
 }


-void vp8_block_variation_c
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int *HVar,
-    int *VVar
-)
-{
-    int i, j;
-    unsigned char *Ptr = src_ptr;
-
-    for (i = 0; i < 4; i++)
-    {
-        for (j = 0; j < 4; j++)
-        {
-            *HVar += abs((int)Ptr[j] - (int)Ptr[j+1]);
-            *VVar += abs((int)Ptr[j] - (int)Ptr[j+src_pixels_per_line]);
-        }
-
-        Ptr += src_pixels_per_line;
-    }
-}
-
-
-
-
 void vp8_sixtap_predict_c
 (
    unsigned char  *src_ptr,
@@ -191,7 +165,7 @@ void vp8_sixtap_predict_c
    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */

-    vp8_filter_block2d(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter);
+    filter_block2d(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter);
 }
 void vp8_sixtap_predict8x8_c
 (
@@ -211,11 +185,11 @@ void vp8_sixtap_predict8x8_c
    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */

    /* First filter 1-D horizontally... */
-    vp8_filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 13, 8, HFilter);
+    filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 13, 8, HFilter);


    /* then filter verticaly... */
-    vp8_filter_block2d_second_pass(FData + 16, dst_ptr, dst_pitch, 8, 8, 8, 8, VFilter);
+    filter_block2d_second_pass(FData + 16, dst_ptr, dst_pitch, 8, 8, 8, 8, VFilter);

 }

@@ -237,11 +211,11 @@ void vp8_sixtap_predict8x4_c
    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */

    /* First filter 1-D horizontally... */
-    vp8_filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 9, 8, HFilter);
+    filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 9, 8, HFilter);


    /* then filter verticaly... */
-    vp8_filter_block2d_second_pass(FData + 16, dst_ptr, dst_pitch, 8, 8, 4, 8, VFilter);
+    filter_block2d_second_pass(FData + 16, dst_ptr, dst_pitch, 8, 8, 4, 8, VFilter);

 }

@@ -264,10 +238,10 @@ void vp8_sixtap_predict16x16_c
    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */

    /* First filter 1-D horizontally... */
-    vp8_filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 21, 16, HFilter);
+    filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 21, 16, HFilter);

    /* then filter verticaly... */
-    vp8_filter_block2d_second_pass(FData + 32, dst_ptr, dst_pitch, 16, 16, 16, 16, VFilter);
+    filter_block2d_second_pass(FData + 32, dst_ptr, dst_pitch, 16, 16, 16, 16, VFilter);

 }

@@ -294,7 +268,7 @@ void vp8_sixtap_predict16x16_c
 *                  Two filter taps should sum to VP8_FILTER_WEIGHT.
 *
 ****************************************************************************/
-void vp8_filter_block2d_bil_first_pass
+static void filter_block2d_bil_first_pass
 (
    unsigned char  *src_ptr,
    unsigned short *dst_ptr,
@@ -345,7 +319,7 @@ void vp8_filter_block2d_bil_first_pass
 *                  Two filter taps should sum to VP8_FILTER_WEIGHT.
 *
 ****************************************************************************/
-void vp8_filter_block2d_bil_second_pass
+static void filter_block2d_bil_second_pass
 (
    unsigned short *src_ptr,
    unsigned char  *dst_ptr,
@@ -399,7 +373,7 @@ void vp8_filter_block2d_bil_second_pass
 *  SPECIAL NOTES : The largest block size can be handled here is 16x16
 *
 ****************************************************************************/
-void vp8_filter_block2d_bil
+static void filter_block2d_bil
 (
    unsigned char *src_ptr,
    unsigned char *dst_ptr,
@@ -415,10 +389,10 @@ void vp8_filter_block2d_bil
    unsigned short FData[17*16];    /* Temp data buffer used in filtering */

    /* First filter 1-D horizontally... */
-    vp8_filter_block2d_bil_first_pass(src_ptr, FData, src_pitch, Height + 1, Width, HFilter);
+    filter_block2d_bil_first_pass(src_ptr, FData, src_pitch, Height + 1, Width, HFilter);

    /* then 1-D vertically... */
-    vp8_filter_block2d_bil_second_pass(FData, dst_ptr, dst_pitch, Height, Width, VFilter);
+    filter_block2d_bil_second_pass(FData, dst_ptr, dst_pitch, Height, Width, VFilter);
 }


@@ -444,19 +418,19 @@ void vp8_bilinear_predict4x4_c
        unsigned char temp2[16];

        bilinear_predict4x4_mmx(src_ptr, src_pixels_per_line, xoffset, yoffset, temp1, 4);
-        vp8_filter_block2d_bil(src_ptr, temp2, src_pixels_per_line, 4, HFilter, VFilter, 4, 4);
+        filter_block2d_bil(src_ptr, temp2, src_pixels_per_line, 4, HFilter, VFilter, 4, 4);

        for (i = 0; i < 16; i++)
        {
            if (temp1[i] != temp2[i])
            {
                bilinear_predict4x4_mmx(src_ptr, src_pixels_per_line, xoffset, yoffset, temp1, 4);
-                vp8_filter_block2d_bil(src_ptr, temp2, src_pixels_per_line, 4, HFilter, VFilter, 4, 4);
+                filter_block2d_bil(src_ptr, temp2, src_pixels_per_line, 4, HFilter, VFilter, 4, 4);
            }
        }
    }
 #endif
-    vp8_filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 4, 4);
+    filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 4, 4);

 }

@@ -476,7 +450,7 @@ void vp8_bilinear_predict8x8_c
    HFilter = vp8_bilinear_filters[xoffset];
    VFilter = vp8_bilinear_filters[yoffset];

-    vp8_filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 8);
+    filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 8);

 }

@@ -496,7 +470,7 @@ void vp8_bilinear_predict8x4_c
    HFilter = vp8_bilinear_filters[xoffset];
    VFilter = vp8_bilinear_filters[yoffset];

-    vp8_filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 4);
+    filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 4);

 }

@@ -516,5 +490,5 @@ void vp8_bilinear_predict16x16_c
    HFilter = vp8_bilinear_filters[xoffset];
    VFilter = vp8_bilinear_filters[yoffset];

-    vp8_filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 16, 16);
+    filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 16, 16);
 }
--- a/vp8/common/findnearmv.c
+++ b/vp8/common/findnearmv.c
@@ -11,6 +11,13 @@

 #include "findnearmv.h"

+const unsigned char vp8_mbsplit_offset[4][16] = {
+    { 0,  8,  0,  0,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
+    { 0,  2,  0,  0,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
+    { 0,  2,  8, 10,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
+    { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15}
+};
+
 /* Predict motion vectors using those from already-decoded nearby blocks.
   Note that we only consider one 4x4 subblock from each candidate 16x16
   macroblock.   */
--- a/vp8/common/findnearmv.h
+++ b/vp8/common/findnearmv.h
@@ -70,4 +70,6 @@ const B_MODE_INFO *vp8_left_bmi(const MODE_INFO *cur_mb, int b);

 const B_MODE_INFO *vp8_above_bmi(const MODE_INFO *cur_mb, int b, int mi_stride);

+extern const unsigned char vp8_mbsplit_offset[4][16];
+
 #endif
--- a/vp8/common/fourcc.hpp
+++ b/vp8/common/fourcc.hpp
@@ -1,121 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef FOURCC_HPP
-#define FOURCC_HPP
-
-#include <iosfwd>
-#include <cstring>
-
-
-#if defined(__POWERPC__) || defined(__APPLE__) || defined(__MERKS__)
-using namespace std;
-#endif
-
-class four_cc
-{
-public:
-
-    four_cc();
-    four_cc(const char*);
-    explicit four_cc(unsigned long);
-
-    bool operator==(const four_cc&) const;
-    bool operator!=(const four_cc&) const;
-
-    bool operator==(const char*) const;
-    bool operator!=(const char*) const;
-
-    operator unsigned long() const;
-    unsigned long as_long() const;
-
-    four_cc& operator=(unsigned long);
-
-    char operator[](int) const;
-
-    std::ostream& put(std::ostream&) const;
-
-    bool printable() const;
-
-private:
-
-    union
-    {
-        char code[4];
-        unsigned long code_as_long;
-    };
-
-};
-
-
-inline four_cc::four_cc()
-{
-}
-
-inline four_cc::four_cc(unsigned long x)
-    : code_as_long(x)
-{
-}
-
-inline four_cc::four_cc(const char* str)
-{
-    memcpy(code, str, 4);
-}
-
-
-inline bool four_cc::operator==(const four_cc& rhs) const
-{
-    return code_as_long == rhs.code_as_long;
-}
-
-inline bool four_cc::operator!=(const four_cc& rhs) const
-{
-    return !operator==(rhs);
-}
-
-inline bool four_cc::operator==(const char* rhs) const
-{
-    return (memcmp(code, rhs, 4) == 0);
-}
-
-inline bool four_cc::operator!=(const char* rhs) const
-{
-    return !operator==(rhs);
-}
-
-
-inline four_cc::operator unsigned long() const
-{
-    return code_as_long;
-}
-
-inline unsigned long four_cc::as_long() const
-{
-    return code_as_long;
-}
-
-inline char four_cc::operator[](int i) const
-{
-    return code[i];
-}
-
-inline four_cc& four_cc::operator=(unsigned long val)
-{
-    code_as_long = val;
-    return *this;
-}
-
-inline std::ostream& operator<<(std::ostream& os, const four_cc& rhs)
-{
-    return rhs.put(os);
-}
-
-#endif
--- a/vp8/common/generic/systemdependent.c
+++ b/vp8/common/generic/systemdependent.c
@@ -10,22 +10,16 @@


 #include "vpx_ports/config.h"
-#include "g_common.h"
-#include "subpixel.h"
-#include "loopfilter.h"
-#include "recon.h"
-#include "idct.h"
-#include "onyxc_int.h"
+#include "vp8/common/g_common.h"
+#include "vp8/common/subpixel.h"
+#include "vp8/common/loopfilter.h"
+#include "vp8/common/recon.h"
+#include "vp8/common/idct.h"
+#include "vp8/common/onyxc_int.h"

 extern void vp8_arch_x86_common_init(VP8_COMMON *ctx);
 extern void vp8_arch_arm_common_init(VP8_COMMON *ctx);

-void (*vp8_build_intra_predictors_mby_ptr)(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby(MACROBLOCKD *x);
-
-void (*vp8_build_intra_predictors_mby_s_ptr)(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x);
-
 void vp8_machine_specific_config(VP8_COMMON *ctx)
 {
 #if CONFIG_RUNTIME_CPU_DETECT
@@ -45,6 +39,10 @@ void vp8_machine_specific_config(VP8_COMMON *ctx)
    rtcd->recon.recon4      = vp8_recon4b_c;
    rtcd->recon.recon_mb    = vp8_recon_mb_c;
    rtcd->recon.recon_mby   = vp8_recon_mby_c;
+    rtcd->recon.build_intra_predictors_mby =
+        vp8_build_intra_predictors_mby;
+    rtcd->recon.build_intra_predictors_mby_s =
+        vp8_build_intra_predictors_mby_s;

    rtcd->subpix.sixtap16x16   = vp8_sixtap_predict16x16_c;
    rtcd->subpix.sixtap8x8     = vp8_sixtap_predict8x8_c;
@@ -75,9 +73,6 @@ void vp8_machine_specific_config(VP8_COMMON *ctx)
 #endif

 #endif
-    /* Pure C: */
-    vp8_build_intra_predictors_mby_ptr = vp8_build_intra_predictors_mby;
-    vp8_build_intra_predictors_mby_s_ptr = vp8_build_intra_predictors_mby_s;

 #if ARCH_X86 || ARCH_X86_64
    vp8_arch_x86_common_init(ctx);
--- a/vp8/common/mbpitch.c
+++ b/vp8/common/mbpitch.c
@@ -17,7 +17,7 @@ typedef enum
    DEST = 1
 } BLOCKSET;

-void vp8_setup_block
+static void setup_block
 (
    BLOCKD *b,
    int mv_stride,
@@ -43,7 +43,8 @@ void vp8_setup_block

 }

-void vp8_setup_macroblock(MACROBLOCKD *x, BLOCKSET bs)
+
+static void setup_macroblock(MACROBLOCKD *x, BLOCKSET bs)
 {
    int block;

@@ -64,16 +65,16 @@ void vp8_setup_macroblock(MACROBLOCKD *x, BLOCKSET bs)

    for (block = 0; block < 16; block++) /* y blocks */
    {
-        vp8_setup_block(&x->block[block], x->dst.y_stride, y, x->dst.y_stride,
+        setup_block(&x->block[block], x->dst.y_stride, y, x->dst.y_stride,
                        (block >> 2) * 4 * x->dst.y_stride + (block & 3) * 4, bs);
    }

    for (block = 16; block < 20; block++) /* U and V blocks */
    {
-        vp8_setup_block(&x->block[block], x->dst.uv_stride, u, x->dst.uv_stride,
+        setup_block(&x->block[block], x->dst.uv_stride, u, x->dst.uv_stride,
                        ((block - 16) >> 1) * 4 * x->dst.uv_stride + (block & 1) * 4, bs);

-        vp8_setup_block(&x->block[block+4], x->dst.uv_stride, v, x->dst.uv_stride,
+        setup_block(&x->block[block+4], x->dst.uv_stride, v, x->dst.uv_stride,
                        ((block - 16) >> 1) * 4 * x->dst.uv_stride + (block & 1) * 4, bs);
    }
 }
@@ -124,6 +125,6 @@ void vp8_build_block_doffsets(MACROBLOCKD *x)
 {

    /* handle the destination pitch features */
-    vp8_setup_macroblock(x, DEST);
-    vp8_setup_macroblock(x, PRED);
+    setup_macroblock(x, DEST);
+    setup_macroblock(x, PRED);
 }
--- a/vp8/common/partialgfupdate.h
+++ b/vp8/common/partialgfupdate.h
@@ -1,19 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef __INC_PARTIALGFUPDATE_H
-#define __INC_PARTIALGFUPDATE_H
-
-#include "onyxc_int.h"
-
-extern void update_gf_selective(ONYX_COMMON *cm, MACROBLOCKD *x);
-
-#endif
--- a/vp8/common/postproc.c
+++ b/vp8/common/postproc.c
@@ -211,7 +211,7 @@ void vp8_post_proc_down_and_across_c
    }
 }

-int vp8_q2mbl(int x)
+static int q2mbl(int x)
 {
    if (x < 20) x = 20;

@@ -314,8 +314,8 @@ static void vp8_deblock_and_de_macro_block(YV12_BUFFER_CONFIG         *source,
    (void) flag;

    POSTPROC_INVOKE(rtcd, downacross)(source->y_buffer, post->y_buffer, source->y_stride,  post->y_stride, source->y_height, source->y_width,  ppl);
-    POSTPROC_INVOKE(rtcd, across)(post->y_buffer, post->y_stride, post->y_height, post->y_width, vp8_q2mbl(q));
-    POSTPROC_INVOKE(rtcd, down)(post->y_buffer, post->y_stride, post->y_height, post->y_width, vp8_q2mbl(q));
+    POSTPROC_INVOKE(rtcd, across)(post->y_buffer, post->y_stride, post->y_height, post->y_width, q2mbl(q));
+    POSTPROC_INVOKE(rtcd, down)(post->y_buffer, post->y_stride, post->y_height, post->y_width, q2mbl(q));

    POSTPROC_INVOKE(rtcd, downacross)(source->u_buffer, post->u_buffer, source->uv_stride, post->uv_stride, source->uv_height, source->uv_width, ppl);
    POSTPROC_INVOKE(rtcd, downacross)(source->v_buffer, post->v_buffer, source->uv_stride, post->uv_stride, source->uv_height, source->uv_width, ppl);
--- a/vp8/common/predictdc.c
+++ b/vp8/common/predictdc.c
@@ -1,44 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#include <stdlib.h>
-#include "blockd.h"
-
-
-void vp8_predict_dc(short *lastdc, short *thisdc, short quant, short *cons)
-{
-    int diff;
-    int sign;
-    int last_dc = *lastdc;
-    int this_dc = *thisdc;
-
-    if (*cons  > DCPREDCNTTHRESH)
-    {
-        this_dc += last_dc;
-    }
-
-    diff = abs(last_dc - this_dc);
-    sign  = (last_dc >> 31) ^(this_dc >> 31);
-    sign |= (!last_dc | !this_dc);
-
-    if (sign)
-    {
-        *cons = 0;
-    }
-    else
-    {
-        if (diff <= DCPREDSIMTHRESH * quant)
-            (*cons)++ ;
-    }
-
-    *thisdc = this_dc;
-    *lastdc = this_dc;
-}
--- a/vp8/common/predictdc.h
+++ b/vp8/common/predictdc.h
@@ -1,18 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef __PREDICTDC_H
-#define __PREDICTDC_H
-
-void uvvp8_predict_dc(short *lastdc, short *thisdc, short quant, short *cons);
-void vp8_predict_dc(short *lastdc, short *thisdc, short quant, short *cons);
-
-#endif
--- a/vp8/common/proposed.h
+++ b/vp8/common/proposed.h
@@ -1,71 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-typedef struct core_codec *codec_ptr;
-typedef struct interface_table *interface_ptr;
-
-typedef struct
-{
-    void (*Initialize)();
-    void (*Shutdown)();
-    codec_ptr(*Create)();
-    int (*compress_frame)(codec_ptr, unsigned int *frame_flags, YV12_BUFFER_CONFIG *sd, unsigned long *size, char *dest, INT64 time_stamp);
-    int (*show_frame)(codec_ptr , YV12_BUFFER_CONFIG *dest, int deblock_level, int noise_level, int flags);
-    void (*Remove)(codec_ptr *comp);
-    interface_ptr(*get_interface)(unsigned int id);
-
-} core_codec;
-
-typedef struct
-{
-    int (*set_bitrate)(codec_ptr, END_USAGE usage, int Datarate);
-    int (*get_bitrate)(codec_ptr, END_USAGE *usage, int *Datarate);
-    int (*set_mode)(codec_ptr, MODE mode, int Speed, char *File);
-    int (*get_mode)(codec_ptr, MODE *mode, int *Speed, char **File);
-} codec_settings_basic;
-
-typedef struct
-{
-    int (*set_bitrate)(codec_ptr, END_USAGE usage, int Datarate);
-    int (*get_bitrate)(codec_ptr, END_USAGE *usage, int *Datarate);
-    int (*set_mode)(codec_ptr, MODE  mode, int  Speed, char   *File);
-    int (*get_mode)(codec_ptr, MODE *mode, int *Speed, char **File);
-    int (*set_denoise)(codec_ptr, int  Level);
-    int (*get_denoise)(codec_ptr, int *Level);
-    int (*set_sharpness)(codec_ptr, int  sharpness);
-    int (*get_sharpness)(codec_ptr, int *sharpness);
-    int (*set_keyframing)(codec_ptr, int  Auto, int  max_distance);
-    int (*get_keyframing)(codec_ptr, int *Auto, int *max_distance);
-    int (*set_buffering)(codec_ptr, int  buffer_level, int  max_buffer_level);
-    int (*get_buffering)(codec_ptr, int *buffer_level, int *max_buffer_level);
-    int (*set_adjust_frame_rate)(codec_ptr, int Allowed, int at_buffer_level_pct);
-    int (*get_adjust_frame_rate)(codec_ptr, int *Allowed, int *at_buffer_level_pct);
-    int (*set_adjust_frame_size)(codec_ptr, int Allowed, int down_at_buffer_level_pct, int up_at_buffer_level_pct);
-    int (*get_adjust_frame_size)(codec_ptr, int *Allowed, int *down_at_buffer_level_pct, int *up_at_buffer_level_pct);
-    int (*set_adjust_quality)(codec_ptr, int Allowed, int min_quantizer, int max_quantizer);
-    int (*get_adjust_quality)(codec_ptr, int *Allowed, int *min_quantizer, int *max_quantizer);
-    int (*set_vbrparms)(codec_ptr, int Bias, int Min, int Max);
-    int (*get_vbrparms)(codec_ptr, int *Bias, int *Min, int *Max);
-
-} codec_settings_v1;
-
-typedef struct
-{
-    int (*request_recovery)(codec_ptr);
-    int (*request_droppable)(codec_ptr);
-    int (*internal_size)(codec_ptr, VPX_SCALING Vertical, VPX_SCALING Horizontal);
-    int (*update_last)(codec_ptr);
-    int (*update_gold)(codec_ptr);
-    int (*use_only_last)(codec_ptr);
-    int (*use_only_gold)(codec_ptr);
-    int (*update_entropy)(codec_ptr);
-
-} codec_realtime_requests;
--- a/vp8/common/recon.h
+++ b/vp8/common/recon.h
@@ -23,6 +23,9 @@
 #define prototype_recon_macroblock(sym) \
    void sym(const struct vp8_recon_rtcd_vtable *rtcd, MACROBLOCKD *x)

+#define prototype_build_intra_predictors(sym) \
+    void sym(MACROBLOCKD *x)
+
 struct vp8_recon_rtcd_vtable;

 #if ARCH_X86 || ARCH_X86_64
@@ -73,9 +76,23 @@ extern prototype_recon_macroblock(vp8_recon_recon_mb);
 #endif
 extern prototype_recon_macroblock(vp8_recon_recon_mby);

+#ifndef vp8_recon_build_intra_predictors_mby
+#define vp8_recon_build_intra_predictors_mby vp8_build_intra_predictors_mby
+#endif
+extern prototype_build_intra_predictors\
+    (vp8_recon_build_intra_predictors_mby);
+
+#ifndef vp8_recon_build_intra_predictors_mby_s
+#define vp8_recon_build_intra_predictors_mby_s vp8_build_intra_predictors_mby_s
+#endif
+extern prototype_build_intra_predictors\
+    (vp8_recon_build_intra_predictors_mby_s);
+
+
 typedef prototype_copy_block((*vp8_copy_block_fn_t));
 typedef prototype_recon_block((*vp8_recon_fn_t));
 typedef prototype_recon_macroblock((*vp8_recon_mb_fn_t));
+typedef prototype_build_intra_predictors((*vp8_build_intra_pred_fn_t));
 typedef struct vp8_recon_rtcd_vtable
 {
    vp8_copy_block_fn_t  copy16x16;
@@ -86,6 +103,8 @@ typedef struct vp8_recon_rtcd_vtable
    vp8_recon_fn_t       recon4;
    vp8_recon_mb_fn_t    recon_mb;
    vp8_recon_mb_fn_t    recon_mby;
+    vp8_build_intra_pred_fn_t  build_intra_predictors_mby_s;
+    vp8_build_intra_pred_fn_t  build_intra_predictors_mby;
 } vp8_recon_rtcd_vtable_t;

 #if CONFIG_RUNTIME_CPU_DETECT
@@ -94,6 +113,5 @@ typedef struct vp8_recon_rtcd_vtable
 #define RECON_INVOKE(ctx,fn) vp8_recon_##fn
 #endif

-void vp8_recon_intra4x4mb(const vp8_recon_rtcd_vtable_t *rtcd, MACROBLOCKD *x);
 void vp8_recon_intra_mbuv(const vp8_recon_rtcd_vtable_t *rtcd, MACROBLOCKD *x);
 #endif
--- a/vp8/common/reconinter.c
+++ b/vp8/common/reconinter.c
@@ -168,7 +168,7 @@ void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, vp8_subpix_fn_t sppf)
    }
 }

-void vp8_build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d, int pitch)
+static void build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d, int pitch)
 {
    unsigned char *ptr_base;
    unsigned char *ptr;
@@ -187,7 +187,7 @@ void vp8_build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d, int pitch)
    }
 }

-void vp8_build_inter_predictors2b(MACROBLOCKD *x, BLOCKD *d, int pitch)
+static void build_inter_predictors2b(MACROBLOCKD *x, BLOCKD *d, int pitch)
 {
    unsigned char *ptr_base;
    unsigned char *ptr;
@@ -246,7 +246,7 @@ void vp8_build_inter_predictors_mbuv(MACROBLOCKD *x)
            BLOCKD *d1 = &x->block[i+1];

            if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-                vp8_build_inter_predictors2b(x, d0, 8);
+                build_inter_predictors2b(x, d0, 8);
            else
            {
                vp8_build_inter_predictors_b(d0, 8, x->subpixel_predict);
@@ -291,7 +291,7 @@ void vp8_build_inter_predictors_mby(MACROBLOCKD *x)
            for (i = 0; i < 4; i++)
            {
                BLOCKD *d = &x->block[bbb[i]];
-                vp8_build_inter_predictors4b(x, d, 16);
+                build_inter_predictors4b(x, d, 16);
            }

        }
@@ -303,7 +303,7 @@ void vp8_build_inter_predictors_mby(MACROBLOCKD *x)
                BLOCKD *d1 = &x->block[i+1];

                if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-                    vp8_build_inter_predictors2b(x, d0, 16);
+                    build_inter_predictors2b(x, d0, 16);
                else
                {
                    vp8_build_inter_predictors_b(d0, 16, x->subpixel_predict);
@@ -372,7 +372,7 @@ void vp8_build_inter_predictors_mb(MACROBLOCKD *x)
            for (i = 0; i < 4; i++)
            {
                BLOCKD *d = &x->block[bbb[i]];
-                vp8_build_inter_predictors4b(x, d, 16);
+                build_inter_predictors4b(x, d, 16);
            }
        }
        else
@@ -383,7 +383,7 @@ void vp8_build_inter_predictors_mb(MACROBLOCKD *x)
                BLOCKD *d1 = &x->block[i+1];

                if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-                    vp8_build_inter_predictors2b(x, d0, 16);
+                    build_inter_predictors2b(x, d0, 16);
                else
                {
                    vp8_build_inter_predictors_b(d0, 16, x->subpixel_predict);
@@ -400,7 +400,7 @@ void vp8_build_inter_predictors_mb(MACROBLOCKD *x)
            BLOCKD *d1 = &x->block[i+1];

            if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-                vp8_build_inter_predictors2b(x, d0, 8);
+                build_inter_predictors2b(x, d0, 8);
            else
            {
                vp8_build_inter_predictors_b(d0, 8, x->subpixel_predict);
@@ -600,7 +600,7 @@ void vp8_build_inter_predictors_mb_s(MACROBLOCKD *x)
            for (i = 0; i < 4; i++)
            {
                BLOCKD *d = &x->block[bbb[i]];
-                /*vp8_build_inter_predictors4b(x, d, 16);*/
+                /*build_inter_predictors4b(x, d, 16);*/

                {
                    unsigned char *ptr_base;
@@ -630,7 +630,7 @@ void vp8_build_inter_predictors_mb_s(MACROBLOCKD *x)

                if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
                {
-                    /*vp8_build_inter_predictors2b(x, d0, 16);*/
+                    /*build_inter_predictors2b(x, d0, 16);*/
                    unsigned char *ptr_base;
                    unsigned char *ptr;
                    unsigned char *pred_ptr = d0->predictor;
@@ -662,7 +662,7 @@ void vp8_build_inter_predictors_mb_s(MACROBLOCKD *x)

            if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
            {
-                /*vp8_build_inter_predictors2b(x, d0, 8);*/
+                /*build_inter_predictors2b(x, d0, 8);*/
                unsigned char *ptr_base;
                unsigned char *ptr;
                unsigned char *pred_ptr = d0->predictor;
--- a/vp8/common/reconintra.h
+++ b/vp8/common/reconintra.h
@@ -14,13 +14,6 @@

 extern void init_intra_left_above_pixels(MACROBLOCKD *x);

-extern void (*vp8_build_intra_predictors_mby_ptr)(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby_neon(MACROBLOCKD *x);
-extern void (*vp8_build_intra_predictors_mby_s_ptr)(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x);
-extern void vp8_build_intra_predictors_mby_s_neon(MACROBLOCKD *x);
-
 extern void vp8_build_intra_predictors_mbuv(MACROBLOCKD *x);
 extern void vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x);

--- a/vp8/common/reconintra4x4.c
+++ b/vp8/common/reconintra4x4.c
@@ -313,89 +313,3 @@ void vp8_intra_prediction_down_copy(MACROBLOCKD *x)
 }


-void vp8_recon_intra4x4mb(const vp8_recon_rtcd_vtable_t *rtcd, MACROBLOCKD *x)
-{
-    int i;
-
-    vp8_intra_prediction_down_copy(x);
-
-#if ARCH_ARM
-    {
-        BLOCKD *b = &x->block[0];
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-        b += 1;
-
-        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-    }
-#else
-    for (i = 0; i < 16; i++)
-    {
-        BLOCKD *b = &x->block[i];
-
-        vp8_predict_intra4x4(b, x->block[i].bmi.mode, x->block[i].predictor);
-        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
-    }
-#endif
-
-    vp8_recon_intra_mbuv(rtcd, x);
-
-}
--- a/vp8/common/threading.h
+++ b/vp8/common/threading.h
@@ -38,6 +38,7 @@
 #define pthread_self() GetCurrentThreadId()
 #else
 #ifdef __APPLE__
+#include <mach/mach_init.h>
 #include <mach/semaphore.h>
 #include <mach/task.h>
 #include <time.h>
--- a/vp8/common/vfwsetting.hpp
+++ b/vp8/common/vfwsetting.hpp
@@ -1,76 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#if !defined(VFWSETTING_HPP)
-#define VFWSETTING_HPP
-//______________________________________________________________________________
-//
-//  VFWSetting.hpp
-//
-
-#include "four_cc.hpp"
-#include <iosfwd>
-
-namespace vpxvp
-{
-
-    //--------------------------------------
-    class VFWSetting
-    {
-        friend std::ostream& operator<<(std::ostream& os, const VFWSetting& vfws);
-
-    public:
-
-        enum Mode
-        {
-            m_setting,
-            m_config
-        };
-
-        enum
-        {
-            header_size = 8,
-            Size = 16
-        };
-
-        VFWSetting(four_cc fcc);
-        ~VFWSetting();
-
-        four_cc fcc() const;
-        Mode mode() const;
-
-        int setting() const;
-        int value() const;
-        void setting_value(int i_setting, int i_value);  //  Sets mode to m_setting
-
-        long size() const;
-        const void* data() const;
-        int data(const void* p_data, unsigned long ul_size);
-
-    private:
-
-        VFWSetting(const VFWSetting& vfws);  //  Not implemented
-        VFWSetting& operator=(const VFWSetting& vfws);  //  Not implemented
-
-        int extract_(const void* p_data, unsigned long ul_size);
-        void update_() const;
-
-        four_cc m_fcc;
-        Mode m_mode;
-        int m_i_setting;
-        int m_i_value;
-
-        mutable unsigned char m_p_data[Size];
-    };
-
-}  //  namespace vpxvp
-
-#endif  //  VFWSETTING_HPP
--- a/vp8/common/vpx_ref_build_prefix.h
+++ b/vp8/common/vpx_ref_build_prefix.h
@@ -1,24 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef _VPX_REF_BUILD_PREFIX_h
-#define _VPX_REF_BUILD_PREFIX_h
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-
-#if defined(__cplusplus)
-}
-#endif
-
-#endif /* include guards */
--- a/vp8/common/vpxblit.h
+++ b/vp8/common/vpxblit.h
@@ -1,112 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef VPXBLIT_H_INCL
-#define VPXBLIT_H_INCL
-/*==============================================================================
-                              Includes
-==============================================================================*/
-
-/*==============================================================================
-                              Defines
-==============================================================================*/
-
-
-#ifdef VPX_BIG_ENDIAN
-#define BYTE_ZERO(X)    ((X & 0xFF000000) >> (24 - 2)   )
-#define BYTE_ONE(X)     ((X & 0x00FF0000) >> (16 - 2)   )
-#define BYTE_TWO(X)     ((X & 0x0000FF00) >> (8 - 2)    )
-#define BYTE_THREE(X)   ((X & 0x000000FF) << (0 + 2)    )
-
-#define BYTE_ZERO_UV(X) ((X & 0x0000FF00) >> (8 - 2)    )
-#define BYTE_ONE_UV(X)  ((X & 0x000000FF) << (0 + 2)    )
-
-#define REREFERENCE(X) (*((int *) &(X)))
-
-#else
-
-#define BYTE_THREE(X)   ((X & 0xFF000000) >> (24 - 2)   )
-#define BYTE_TWO(X)     ((X & 0x00FF0000) >> (16 - 2)   )
-#define BYTE_ONE(X)     ((X & 0x0000FF00) >> (8 - 2)    )
-#define BYTE_ZERO(X)    ((X & 0x000000FF) << (0 + 2)    )
-
-#define BYTE_ONE_UV(X) ((X & 0x0000FF00) >> (8 - 2) )
-#define BYTE_ZERO_UV(X)     ((X & 0x000000FF) << (0 + 2)    )
-
-#define REREFERENCE(X) (*((int *) &(X)))
-
-#endif
-
-
-/*==============================================================================
-                            Type Definitions
-==============================================================================*/
-typedef struct  // YUV buffer configuration structure
-{
-    int   y_width;
-    int   y_height;
-    int   y_stride;
-
-    int   uv_width;
-    int   uv_height;
-    int   uv_stride;
-
-    char *y_buffer;
-    char *u_buffer;
-    char *v_buffer;
-
-    char *uv_start;
-    int   uv_dst_area;
-    int   uv_used_area;
-
-} VPX_BLIT_CONFIG;
-
-typedef struct tx86_params
-{
-    unsigned int pushed_registers[6];
-    unsigned int return_address;
-    unsigned int dst;
-    unsigned int scrn_pitch;
-    VPX_BLIT_CONFIG *buff_config;
-} x86_params;
-
-/*=============================================================================
-                                Enums
-==============================================================================*/
-
-
-/*==============================================================================
-                              Structures
-==============================================================================*/
-
-/*==============================================================================
-                             Constants
-==============================================================================*/
-
-
-/*==============================================================================
-                               Variables
-==============================================================================*/
-
-
-
-
-/*==============================================================================
-                            Function Protoypes/MICROS
-==============================================================================*/
-int vpx_get_size_of_pixel(unsigned int bd);
-void *vpx_get_blitter(unsigned int bd);
-void vpx_set_blit(void);
-void vpx_destroy_blit(void);
-
-
-
-#endif //VPXBLIT_H_INCL
--- a/vp8/common/vpxblit_c64.h
+++ b/vp8/common/vpxblit_c64.h
@@ -1,48 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef _VPX_BLIT_C64_h
-#define _VPX_BLIT_C64_h
-
-/****************************************************************************
-*  Typedefs
-****************************************************************************/
-
-typedef struct  // YUV buffer configuration structure
-{
-    int     y_width;
-    int     y_height;
-    int     y_stride;
-
-    int     uv_width;
-    int     uv_height;
-    int     uv_stride;
-
-    unsigned char *y_buffer;
-    unsigned char *u_buffer;
-    unsigned char *v_buffer;
-
-    unsigned char *y_ptr_scrn;
-    unsigned char *u_ptr_scrn;
-    unsigned char *v_ptr_scrn;
-
-} DXV_YUV_BUFFER_CONFIG;
-
-typedef struct
-{
-    unsigned char *rgbptr_scrn;
-    unsigned char *y_ptr_scrn;
-    unsigned char *u_ptr_scrn;
-    unsigned char *v_ptr_scrn;
-    unsigned char *rgbptr_scrn2;
-} DXV_FINAL_VIDEO;
-
-#endif /* include guards */
--- a/vp8/common/vpxerrors.h
+++ b/vp8/common/vpxerrors.h
@@ -1,13 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-
-#define ALLOC_FAILURE -2
--- a/vp8/common/x86/loopfilter_x86.c
+++ b/vp8/common/x86/loopfilter_x86.c
@@ -10,7 +10,7 @@


 #include "vpx_ports/config.h"
-#include "loopfilter.h"
+#include "vp8/common/loopfilter.h"

 prototype_loopfilter(vp8_loop_filter_horizontal_edge_c);
 prototype_loopfilter(vp8_loop_filter_vertical_edge_c);
--- a/vp8/common/x86/subpixel_mmx.asm
+++ b/vp8/common/x86/subpixel_mmx.asm
@@ -113,97 +113,6 @@ nextrow:
    ret


-;
-; THIS FUNCTION APPEARS TO BE UNUSED
-;
-;void vp8_filter_block1d_v6_mmx
-;(
-;   short *src_ptr,
-;   unsigned char *output_ptr,
-;   unsigned int pixels_per_line,
-;   unsigned int pixel_step,
-;   unsigned int output_height,
-;   unsigned int output_width,
-;   short * vp8_filter
-;)
-global sym(vp8_filter_block1d_v6_mmx)
-sym(vp8_filter_block1d_v6_mmx):
-    push        rbp
-    mov         rbp, rsp
-    SHADOW_ARGS_TO_STACK 7
-    GET_GOT     rbx
-    push        rsi
-    push        rdi
-    ; end prolog
-
-        movq      mm5, [GLOBAL(rd)]
-        push        rbx
-        mov         rbx, arg(6) ;vp8_filter
-        movq      mm1, [rbx + 16]             ; do both the negative taps first!!!
-        movq      mm2, [rbx + 32]         ;
-        movq      mm6, [rbx + 48]        ;
-        movq      mm7, [rbx + 64]        ;
-
-        movsxd      rdx, dword ptr arg(2) ;pixels_per_line
-        mov         rdi, arg(1) ;output_ptr
-        mov         rsi, arg(0) ;src_ptr
-        sub         rsi, rdx
-        sub         rsi, rdx
-        movsxd      rcx, DWORD PTR arg(4) ;output_height
-        movsxd      rax, DWORD PTR arg(5) ;output_width      ; destination pitch?
-        pxor        mm0, mm0              ; mm0 = 00000000
-
-
-nextrow_v:
-        movq        mm3, [rsi+rdx]        ; mm3 = p0..p8  = row -1
-        pmullw      mm3, mm1              ; mm3 *= kernel 1 modifiers.
-
-
-        movq        mm4, [rsi + 4*rdx]      ; mm4 = p0..p3  = row 2
-        pmullw      mm4, mm7              ; mm4 *= kernel 4 modifiers.
-        paddsw      mm3, mm4              ; mm3 += mm4
-
-        movq        mm4, [rsi + 2*rdx]           ; mm4 = p0..p3  = row 0
-        pmullw      mm4, mm2              ; mm4 *= kernel 2 modifiers.
-        paddsw      mm3, mm4              ; mm3 += mm4
-
-        movq        mm4, [rsi]            ; mm4 = p0..p3  = row -2
-        pmullw      mm4, [rbx]            ; mm4 *= kernel 0 modifiers.
-        paddsw      mm3, mm4              ; mm3 += mm4
-
-
-        add         rsi, rdx              ; move source forward 1 line to avoid 3 * pitch
-        movq        mm4, [rsi + 2*rdx]     ; mm4 = p0..p3  = row 1
-        pmullw      mm4, mm6              ; mm4 *= kernel 3 modifiers.
-        paddsw      mm3, mm4              ; mm3 += mm4
-
-        movq        mm4, [rsi + 4*rdx]    ; mm4 = p0..p3  = row 3
-        pmullw      mm4, [rbx +80]        ; mm4 *= kernel 3 modifiers.
-        paddsw      mm3, mm4              ; mm3 += mm4
-
-
-        paddsw      mm3, mm5               ; mm3 += round value
-        psraw       mm3, VP8_FILTER_SHIFT     ; mm3 /= 128
-        packuswb    mm3, mm0              ; pack and saturate
-
-        movd        [rdi],mm3             ; store the results in the destination
-
-        add         rdi,rax;
-
-        dec         rcx                   ; decrement count
-        jnz         nextrow_v             ; next row
-
-        pop         rbx
-
-    ; begin epilog
-    pop rdi
-    pop rsi
-    RESTORE_GOT
-    UNSHADOW_ARGS
-    pop         rbp
-    ret
-
-
 ;void vp8_filter_block1dc_v6_mmx
 ;(
 ;   short *src_ptr,
--- a/vp8/common/x86/vp8_asm_stubs.c
+++ b/vp8/common/x86/vp8_asm_stubs.c
@@ -11,7 +11,7 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/mem.h"
-#include "subpixel.h"
+#include "vp8/common/subpixel.h"

 extern const short vp8_six_tap_mmx[8][6*8];
 extern const short vp8_bilinear_filters_mmx[8][2*8];
--- a/vp8/common/x86/x86_systemdependent.c
+++ b/vp8/common/x86/x86_systemdependent.c
@@ -11,13 +11,13 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/x86.h"
-#include "g_common.h"
-#include "subpixel.h"
-#include "loopfilter.h"
-#include "recon.h"
-#include "idct.h"
-#include "pragmas.h"
-#include "onyxc_int.h"
+#include "vp8/common/g_common.h"
+#include "vp8/common/subpixel.h"
+#include "vp8/common/loopfilter.h"
+#include "vp8/common/recon.h"
+#include "vp8/common/idct.h"
+#include "vp8/common/pragmas.h"
+#include "vp8/common/onyxc_int.h"

 void vp8_arch_x86_common_init(VP8_COMMON *ctx)
 {
--- a/vp8/decoder/arm/arm_dsystemdependent.c
+++ b/vp8/decoder/arm/arm_dsystemdependent.c
@@ -11,11 +11,11 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/arm.h"
-#include "blockd.h"
-#include "pragmas.h"
-#include "postproc.h"
-#include "dequantize.h"
-#include "onyxd_int.h"
+#include "vp8/common/blockd.h"
+#include "vp8/common/pragmas.h"
+#include "vp8/common/postproc.h"
+#include "vp8/decoder/dequantize.h"
+#include "vp8/decoder/onyxd_int.h"

 void vp8_arch_arm_decode_init(VP8D_COMP *pbi)
 {
--- a/vp8/decoder/arm/armv6/idct_blk_v6.c
+++ b/vp8/decoder/arm/armv6/idct_blk_v6.c
@@ -9,8 +9,8 @@
 */

 #include "vpx_ports/config.h"
-#include "idct.h"
-#include "dequantize.h"
+#include "vp8/common/idct.h"
+#include "vp8/decoder/dequantize.h"

 void vp8_dequant_dc_idct_add_y_block_v6
            (short *q, short *dq, unsigned char *pre,
--- a/vp8/decoder/arm/dequantize_arm.c
+++ b/vp8/decoder/arm/dequantize_arm.c
@@ -10,9 +10,8 @@


 #include "vpx_ports/config.h"
-#include "dequantize.h"
-#include "predictdc.h"
-#include "idct.h"
+#include "vp8/decoder/dequantize.h"
+#include "vp8/common/idct.h"
 #include "vpx_mem/vpx_mem.h"

 #if HAVE_ARMV7
--- a/vp8/decoder/arm/neon/idct_blk_neon.c
+++ b/vp8/decoder/arm/neon/idct_blk_neon.c
@@ -9,8 +9,8 @@
 */

 #include "vpx_ports/config.h"
-#include "idct.h"
-#include "dequantize.h"
+#include "vp8/common/idct.h"
+#include "vp8/decoder/dequantize.h"

 /* place these declarations here because we don't want to maintain them
 * outside of this scope
--- a/vp8/decoder/decodemv.c
+++ b/vp8/decoder/decodemv.c
@@ -10,10 +10,10 @@


 #include "treereader.h"
-#include "entropymv.h"
-#include "entropymode.h"
+#include "vp8/common/entropymv.h"
+#include "vp8/common/entropymode.h"
 #include "onyxd_int.h"
-#include "findnearmv.h"
+#include "vp8/common/findnearmv.h"

 #if CONFIG_DEBUG
 #include <assert.h>
@@ -228,15 +228,8 @@ unsigned int vp8_mv_cont_count[5][4] =
 };
 #endif

-unsigned char vp8_mbsplit_offset[4][16] = {
-    { 0,  8,  0,  0,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
-    { 0,  2,  0,  0,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
-    { 0,  2,  8, 10,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
-    { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15}
-};
-
-unsigned char vp8_mbsplit_fill_count[4] = {8, 8, 4, 1};
-unsigned char vp8_mbsplit_fill_offset[4][16] = {
+static const unsigned char mbsplit_fill_count[4] = {8, 8, 4, 1};
+static const unsigned char mbsplit_fill_offset[4][16] = {
    { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15},
    { 0,  1,  4,  5,  8,  9, 12, 13,  2,  3,   6,  7, 10, 11, 14, 15},
    { 0,  1,  4,  5,  2,  3,  6,  7,  8,  9,  12, 13, 10, 11, 14, 15},
@@ -246,7 +239,7 @@ unsigned char vp8_mbsplit_fill_offset[4][16] = {



-void vp8_mb_mode_mv_init(VP8D_COMP *pbi)
+static void mb_mode_mv_init(VP8D_COMP *pbi)
 {
    vp8_reader *const bc = & pbi->bc;
    MV_CONTEXT *const mvc = pbi->common.fc.mvc;
@@ -287,7 +280,7 @@ void vp8_mb_mode_mv_init(VP8D_COMP *pbi)
    }
 }

-void vp8_read_mb_modes_mv(VP8D_COMP *pbi, MODE_INFO *mi, MB_MODE_INFO *mbmi,
+static void read_mb_modes_mv(VP8D_COMP *pbi, MODE_INFO *mi, MB_MODE_INFO *mbmi,
                            int mb_row, int mb_col)
 {
    const MV Zero = { 0, 0};
@@ -405,10 +398,10 @@ void vp8_read_mb_modes_mv(VP8D_COMP *pbi, MODE_INFO *mi, MB_MODE_INFO *mbmi,
                    /* Fill (uniform) modes, mvs of jth subset.
                     Must do it here because ensuing subsets can
                     refer back to us via "left" or "above". */
-                    unsigned char *fill_offset;
-                    unsigned int fill_count = vp8_mbsplit_fill_count[s];
+                    const unsigned char *fill_offset;
+                    unsigned int fill_count = mbsplit_fill_count[s];

-                    fill_offset = &vp8_mbsplit_fill_offset[s][(unsigned char)j * vp8_mbsplit_fill_count[s]];
+                    fill_offset = &mbsplit_fill_offset[s][(unsigned char)j * mbsplit_fill_count[s]];

                    do {
                        mi->bmi[ *fill_offset] = bmi;
@@ -525,7 +518,7 @@ void vp8_decode_mode_mvs(VP8D_COMP *pbi)
    MODE_INFO *mi = pbi->common.mi;
    int mb_row = -1;

-    vp8_mb_mode_mv_init(pbi);
+    mb_mode_mv_init(pbi);

    while (++mb_row < pbi->common.mb_rows)
    {
@@ -543,11 +536,11 @@ void vp8_decode_mode_mvs(VP8D_COMP *pbi)

        while (++mb_col < pbi->common.mb_cols)
        {
-            /*vp8_read_mb_modes_mv(pbi, xd->mode_info_context, &xd->mode_info_context->mbmi, mb_row, mb_col);*/
+            /*read_mb_modes_mv(pbi, xd->mode_info_context, &xd->mode_info_context->mbmi, mb_row, mb_col);*/
            if(pbi->common.frame_type == KEY_FRAME)
                vp8_kfread_modes(pbi, mi, mb_row, mb_col);
            else
-                vp8_read_mb_modes_mv(pbi, mi, &mi->mbmi, mb_row, mb_col);
+                read_mb_modes_mv(pbi, mi, &mi->mbmi, mb_row, mb_col);

            mi++;       /* next macroblock */
        }
--- a/vp8/decoder/decodframe.c
+++ b/vp8/decoder/decodframe.c
@@ -10,28 +10,27 @@


 #include "onyxd_int.h"
-#include "header.h"
-#include "reconintra.h"
-#include "reconintra4x4.h"
-#include "recon.h"
-#include "reconinter.h"
+#include "vp8/common/header.h"
+#include "vp8/common/reconintra.h"
+#include "vp8/common/reconintra4x4.h"
+#include "vp8/common/recon.h"
+#include "vp8/common/reconinter.h"
 #include "dequantize.h"
 #include "detokenize.h"
-#include "invtrans.h"
-#include "alloccommon.h"
-#include "entropymode.h"
-#include "quant_common.h"
+#include "vp8/common/invtrans.h"
+#include "vp8/common/alloccommon.h"
+#include "vp8/common/entropymode.h"
+#include "vp8/common/quant_common.h"
 #include "vpx_scale/vpxscale.h"
 #include "vpx_scale/yv12extend.h"
-#include "setupintrarecon.h"
+#include "vp8/common/setupintrarecon.h"

 #include "decodemv.h"
-#include "extend.h"
+#include "vp8/common/extend.h"
 #include "vpx_mem/vpx_mem.h"
-#include "idct.h"
+#include "vp8/common/idct.h"
 #include "dequantize.h"
-#include "predictdc.h"
-#include "threading.h"
+#include "vp8/common/threading.h"
 #include "decoderthreading.h"
 #include "dboolhuff.h"

@@ -116,8 +115,8 @@ static void skip_recon_mb(VP8D_COMP *pbi, MACROBLOCKD *xd)
    {

        vp8_build_intra_predictors_mbuv_s(xd);
-        vp8_build_intra_predictors_mby_s_ptr(xd);
-
+        RECON_INVOKE(&pbi->common.rtcd.recon,
+                     build_intra_predictors_mby_s)(xd);
    }
    else
    {
@@ -176,7 +175,7 @@ void clamp_mvs(MACROBLOCKD *xd)

 }

-void vp8_decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd)
+static void decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd)
 {
    int eobtotal = 0;
    int i, do_clamp = xd->mode_info_context->mbmi.need_to_clamp_mvs;
@@ -215,7 +214,8 @@ void vp8_decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd)

        if (xd->mode_info_context->mbmi.mode != B_PRED)
        {
-            vp8_build_intra_predictors_mby_ptr(xd);
+            RECON_INVOKE(&pbi->common.rtcd.recon,
+                         build_intra_predictors_mby)(xd);
        } else {
            vp8_intra_prediction_down_copy(xd);
        }
@@ -320,10 +320,8 @@ FILE *vpxlog = 0;



-void vp8_decode_mb_row(VP8D_COMP *pbi,
-                       VP8_COMMON *pc,
-                       int mb_row,
-                       MACROBLOCKD *xd)
+static void
+decode_mb_row(VP8D_COMP *pbi, VP8_COMMON *pc, int mb_row, MACROBLOCKD *xd)
 {

    int i;
@@ -395,7 +393,7 @@ void vp8_decode_mb_row(VP8D_COMP *pbi,
        else
        pbi->debugoutput =0;
        */
-        vp8_decode_macroblock(pbi, xd);
+        decode_macroblock(pbi, xd);

        /* check if the boolean decoder has suffered an error */
        xd->corrupted |= vp8dx_bool_error(xd->current_bc);
@@ -901,7 +899,7 @@ int vp8_decode_frame(VP8D_COMP *pbi)
                    ibc = 0;
            }

-            vp8_decode_mb_row(pbi, pc, mb_row, xd);
+            decode_mb_row(pbi, pc, mb_row, xd);
        }
    }

--- a/vp8/decoder/dequantize.c
+++ b/vp8/decoder/dequantize.c
@@ -11,8 +11,7 @@

 #include "vpx_ports/config.h"
 #include "dequantize.h"
-#include "predictdc.h"
-#include "idct.h"
+#include "vp8/common/idct.h"
 #include "vpx_mem/vpx_mem.h"

 extern void vp8_short_idct4x4llm_c(short *input, short *output, int pitch) ;
--- a/vp8/decoder/dequantize.h
+++ b/vp8/decoder/dequantize.h
@@ -11,7 +11,7 @@

 #ifndef DEQUANTIZE_H
 #define DEQUANTIZE_H
-#include "blockd.h"
+#include "vp8/common/blockd.h"

 #define prototype_dequant_block(sym) \
    void sym(BLOCKD *x)
--- a/vp8/decoder/detokenize.c
+++ b/vp8/decoder/detokenize.c
@@ -9,8 +9,8 @@
 */


-#include "type_aliases.h"
-#include "blockd.h"
+#include "vp8/common/type_aliases.h"
+#include "vp8/common/blockd.h"
 #include "onyxd_int.h"
 #include "vpx_mem/vpx_mem.h"
 #include "vpx_ports/mem.h"
@@ -19,7 +19,13 @@
 #define BOOL_DATA UINT8

 #define OCB_X PREV_COEF_CONTEXTS * ENTROPY_NODES
-DECLARE_ALIGNED(16, UINT8, vp8_coef_bands_x[16]) = { 0, 1 * OCB_X, 2 * OCB_X, 3 * OCB_X, 6 * OCB_X, 4 * OCB_X, 5 * OCB_X, 6 * OCB_X, 6 * OCB_X, 6 * OCB_X, 6 * OCB_X, 6 * OCB_X, 6 * OCB_X, 6 * OCB_X, 6 * OCB_X, 7 * OCB_X};
+DECLARE_ALIGNED(16, static const unsigned char, coef_bands_x[16]) =
+{
+    0 * OCB_X, 1 * OCB_X, 2 * OCB_X, 3 * OCB_X,
+    6 * OCB_X, 4 * OCB_X, 5 * OCB_X, 6 * OCB_X,
+    6 * OCB_X, 6 * OCB_X, 6 * OCB_X, 6 * OCB_X,
+    6 * OCB_X, 6 * OCB_X, 6 * OCB_X, 7 * OCB_X
+};
 #define EOB_CONTEXT_NODE            0
 #define ZERO_CONTEXT_NODE           1
 #define ONE_CONTEXT_NODE            2
@@ -135,7 +141,7 @@ DECLARE_ALIGNED(16, extern const unsigned char, vp8dx_bitreader_norm[256]);
            Prob = coef_probs; \
            if(c<15) {\
            ++c; \
-            Prob += vp8_coef_bands_x[c]; \
+            Prob += coef_bands_x[c]; \
            goto branch; \
            } goto BLOCK_FINISHED; /*for malformed input */\
        } \
@@ -244,7 +250,7 @@ BLOCK_LOOP:
    Prob += v * ENTROPY_NODES;

 DO_WHILE:
-    Prob += vp8_coef_bands_x[c];
+    Prob += coef_bands_x[c];
    DECODE_AND_BRANCH_IF_ZERO(Prob[EOB_CONTEXT_NODE], BLOCK_FINISHED);

 CHECK_0_:
--- a/vp8/decoder/generic/dsystemdependent.c
+++ b/vp8/decoder/generic/dsystemdependent.c
@@ -10,8 +10,8 @@


 #include "vpx_ports/config.h"
-#include "dequantize.h"
-#include "onyxd_int.h"
+#include "vp8/decoder/dequantize.h"
+#include "vp8/decoder/onyxd_int.h"

 extern void vp8_arch_x86_decode_init(VP8D_COMP *pbi);
 extern void vp8_arch_arm_decode_init(VP8D_COMP *pbi);
--- a/vp8/decoder/idct_blk.c
+++ b/vp8/decoder/idct_blk.c
@@ -9,7 +9,7 @@
 */

 #include "vpx_ports/config.h"
-#include "idct.h"
+#include "vp8/common/idct.h"
 #include "dequantize.h"

 void vp8_dequant_dc_idct_add_c(short *input, short *dq, unsigned char *pred,
--- a/vp8/decoder/onyxd_if.c
+++ b/vp8/decoder/onyxd_if.c
@@ -9,25 +9,25 @@
 */


-#include "onyxc_int.h"
+#include "vp8/common/onyxc_int.h"
 #if CONFIG_POSTPROC
-#include "postproc.h"
+#include "vp8/common/postproc.h"
 #endif
-#include "onyxd.h"
+#include "vp8/common/onyxd.h"
 #include "onyxd_int.h"
 #include "vpx_mem/vpx_mem.h"
-#include "alloccommon.h"
+#include "vp8/common/alloccommon.h"
 #include "vpx_scale/yv12extend.h"
-#include "loopfilter.h"
-#include "swapyv12buffer.h"
-#include "g_common.h"
-#include "threading.h"
+#include "vp8/common/loopfilter.h"
+#include "vp8/common/swapyv12buffer.h"
+#include "vp8/common/g_common.h"
+#include "vp8/common/threading.h"
 #include "decoderthreading.h"
 #include <stdio.h>

-#include "quant_common.h"
+#include "vp8/common/quant_common.h"
 #include "vpx_scale/vpxscale.h"
-#include "systemdependent.h"
+#include "vp8/common/systemdependent.h"
 #include "vpx_ports/vpx_timer.h"
 #include "detokenize.h"
 #if ARCH_ARM
@@ -37,43 +37,6 @@
 extern void vp8_init_loop_filter(VP8_COMMON *cm);
 extern void vp8cx_init_de_quantizer(VP8D_COMP *pbi);

-#if CONFIG_DEBUG
-void vp8_recon_write_yuv_frame(unsigned char *name, YV12_BUFFER_CONFIG *s)
-{
-    FILE *yuv_file = fopen((char *)name, "ab");
-    unsigned char *src = s->y_buffer;
-    int h = s->y_height;
-
-    do
-    {
-        fwrite(src, s->y_width, 1,  yuv_file);
-        src += s->y_stride;
-    }
-    while (--h);
-
-    src = s->u_buffer;
-    h = s->uv_height;
-
-    do
-    {
-        fwrite(src, s->uv_width, 1,  yuv_file);
-        src += s->uv_stride;
-    }
-    while (--h);
-
-    src = s->v_buffer;
-    h = s->uv_height;
-
-    do
-    {
-        fwrite(src, s->uv_width, 1, yuv_file);
-        src += s->uv_stride;
-    }
-    while (--h);
-
-    fclose(yuv_file);
-}
-#endif

 void vp8dx_initialize()
 {
@@ -155,35 +118,6 @@ void vp8dx_remove_decompressor(VP8D_PTR ptr)
 }


-void vp8dx_set_setting(VP8D_PTR comp, VP8D_SETTING oxst, int x)
-{
-    VP8D_COMP *pbi = (VP8D_COMP *) comp;
-
-    (void) pbi;
-    (void) x;
-
-    switch (oxst)
-    {
-    case VP8D_OK:
-        break;
-    }
-}
-
-int vp8dx_get_setting(VP8D_PTR comp, VP8D_SETTING oxst)
-{
-    VP8D_COMP *pbi = (VP8D_COMP *) comp;
-
-    (void) pbi;
-
-    switch (oxst)
-    {
-    case VP8D_OK:
-        break;
-    }
-
-    return -1;
-}
-
 int vp8dx_get_reference(VP8D_PTR ptr, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd)
 {
    VP8D_COMP *pbi = (VP8D_COMP *) ptr;
@@ -203,6 +137,8 @@ int vp8dx_get_reference(VP8D_PTR ptr, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_C

    return 0;
 }
+
+
 int vp8dx_set_reference(VP8D_PTR ptr, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd)
 {
    VP8D_COMP *pbi = (VP8D_COMP *) ptr;
@@ -459,12 +395,6 @@ int vp8dx_receive_compressed_data(VP8D_PTR ptr, unsigned long size, const unsign
        vp8_yv12_extend_frame_borders_ptr(cm->frame_to_show);
    }

-#if 0
-    /* DEBUG code */
-    /*vp8_recon_write_yuv_frame("recon.yuv", cm->frame_to_show);*/
-    if (cm->current_video_frame <= 5)
-        write_dx_frame_to_file(cm->frame_to_show, cm->current_video_frame);
-#endif

    vp8_clear_system_state();

--- a/vp8/decoder/onyxd_int.h
+++ b/vp8/decoder/onyxd_int.h
@@ -12,10 +12,10 @@
 #ifndef __INC_VP8D_INT_H
 #define __INC_VP8D_INT_H
 #include "vpx_ports/config.h"
-#include "onyxd.h"
+#include "vp8/common/onyxd.h"
 #include "treereader.h"
-#include "onyxc_int.h"
-#include "threading.h"
+#include "vp8/common/onyxc_int.h"
+#include "vp8/common/threading.h"
 #include "dequantize.h"

 typedef struct
@@ -113,6 +113,7 @@ typedef struct VP8Decompressor
    pthread_t           *h_decoding_thread;
    sem_t               *h_event_start_decoding;
    sem_t                h_event_end_decoding;
+    sem_t				*h_mb_counter;
    /* end of threading data */
 #endif

--- a/vp8/decoder/reconintra_mt.c
+++ b/vp8/decoder/reconintra_mt.c
@@ -10,8 +10,8 @@


 #include "vpx_ports/config.h"
-#include "recon.h"
-#include "reconintra.h"
+#include "vp8/common/recon.h"
+#include "vp8/common/reconintra.h"
 #include "vpx_mem/vpx_mem.h"
 #include "onyxd_int.h"

--- a/vp8/decoder/threading.c
+++ b/vp8/decoder/threading.c
@@ -12,18 +12,15 @@
 #if !defined(WIN32) && CONFIG_OS_SUPPORT == 1
 # include <unistd.h>
 #endif
-#ifdef __APPLE__
-#include <mach/mach_init.h>
-#endif
 #include "onyxd_int.h"
 #include "vpx_mem/vpx_mem.h"
-#include "threading.h"
+#include "vp8/common/threading.h"

-#include "loopfilter.h"
-#include "extend.h"
+#include "vp8/common/loopfilter.h"
+#include "vp8/common/extend.h"
 #include "vpx_ports/vpx_timer.h"
 #include "detokenize.h"
-#include "reconinter.h"
+#include "vp8/common/reconinter.h"
 #include "reconintra_mt.h"

 extern void mb_init_dequantizer(VP8D_COMP *pbi, MACROBLOCKD *xd);
@@ -36,7 +33,7 @@ extern void vp8_build_uvmvs(MACROBLOCKD *x, int fullpixel);
 #define RTCD_VTABLE(x) NULL
 #endif

-void vp8_setup_decoding_thread_data(VP8D_COMP *pbi, MACROBLOCKD *xd, MB_ROW_DEC *mbrd, int count)
+static void setup_decoding_thread_data(VP8D_COMP *pbi, MACROBLOCKD *xd, MB_ROW_DEC *mbrd, int count)
 {
    VP8_COMMON *const pc = & pbi->common;
    int i, j;
@@ -90,7 +87,7 @@ void vp8_setup_decoding_thread_data(VP8D_COMP *pbi, MACROBLOCKD *xd, MB_ROW_DEC
 }


-void vp8mt_decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd, int mb_row, int mb_col)
+static void decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd, int mb_row, int mb_col)
 {
    int eobtotal = 0;
    int i, do_clamp = xd->mode_info_context->mbmi.need_to_clamp_mvs;
@@ -217,7 +214,7 @@ void vp8mt_decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd, int mb_row, int mb
 }


-THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)
+static THREAD_FUNCTION thread_decoding_proc(void *p_data)
 {
    int ithread = ((DECODETHREAD_DATA *)p_data)->ithread;
    VP8D_COMP *pbi = (VP8D_COMP *)(((DECODETHREAD_DATA *)p_data)->ptr1);
@@ -278,6 +275,7 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)

                    for (mb_col = 0; mb_col < pc->mb_cols; mb_col++)
                    {
+/*
                        if ((mb_col & (nsync-1)) == 0)
                        {
                            while (mb_col > (*last_row_current_mb_col - nsync) && *last_row_current_mb_col != pc->mb_cols - 1)
@@ -286,6 +284,8 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)
                                thread_sleep(0);
                            }
                        }
+*/
+                        sem_wait(&pbi->h_mb_counter[ithread]);

                        if (xd->mode_info_context->mbmi.mode == SPLITMV || xd->mode_info_context->mbmi.mode == B_PRED)
                        {
@@ -296,18 +296,6 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)
                            }
                        }

-                        if(pbi->common.filter_level)
-                        {
-                            /*update loopfilter info*/
-                            Segment = (alt_flt_enabled) ? xd->mode_info_context->mbmi.segment_id : 0;
-                            filter_level = pbi->mt_baseline_filter_level[Segment];
-                            /* Distance of Mb to the various image edges.
-                             * These are specified to 8th pel as they are always compared to values that are in 1/8th pel units
-                             * Apply any context driven MB level adjustment
-                             */
-                            filter_level = vp8_adjust_mb_lf_value(xd, filter_level);
-                        }
-
                        /* Distance of Mb to the various image edges.
                         * These are specified to 8th pel as they are always compared to values that are in 1/8th pel units
                         */
@@ -333,7 +321,7 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)
                        xd->pre.v_buffer = pc->yv12_fb[ref_fb_idx].v_buffer + recon_uvoffset;

                        vp8_build_uvmvs(xd, pc->full_pixel);
-                        vp8mt_decode_macroblock(pbi, xd, mb_row, mb_col);
+                        decode_macroblock(pbi, xd, mb_row, mb_col);

                        if (pbi->common.filter_level)
                        {
@@ -362,7 +350,16 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)
                                }
                            }

-                          /* loopfilter on this macroblock. */
+                            /* update loopfilter info */
+                            Segment = (alt_flt_enabled) ? xd->mode_info_context->mbmi.segment_id : 0;
+                            filter_level = pbi->mt_baseline_filter_level[Segment];
+                            /* Distance of Mb to the various image edges.
+                             * These are specified to 8th pel as they are always compared to values that are in 1/8th pel units
+                             * Apply any context driven MB level adjustment
+                             */
+                            filter_level = vp8_adjust_mb_lf_value(xd, filter_level);
+
+                            /* loopfilter on this macroblock. */
                            if (filter_level)
                            {
                                if (mb_col > 0)
@@ -389,6 +386,9 @@ THREAD_FUNCTION vp8_thread_decoding_proc(void *p_data)

                        /*pbi->mb_row_di[ithread].current_mb_col = mb_col;*/
                        pbi->mt_current_mb_col[mb_row] = mb_col;
+
+                        if (mb_row != pbi->common.mb_rows-1)
+                            sem_post(&pbi->h_mb_counter[ithread+1]);
                    }

                    /* adjust to the next row of mbs */
@@ -444,6 +444,7 @@ void vp8_decoder_create_threads(VP8D_COMP *pbi)

        CHECK_MEM_ERROR(pbi->h_decoding_thread, vpx_malloc(sizeof(pthread_t) * pbi->decoding_thread_count));
        CHECK_MEM_ERROR(pbi->h_event_start_decoding, vpx_malloc(sizeof(sem_t) * pbi->decoding_thread_count));
+        CHECK_MEM_ERROR(pbi->h_mb_counter, vpx_malloc(sizeof(sem_t) * (pbi->decoding_thread_count + 1)));
        CHECK_MEM_ERROR(pbi->mb_row_di, vpx_memalign(32, sizeof(MB_ROW_DEC) * pbi->decoding_thread_count));
        vpx_memset(pbi->mb_row_di, 0, sizeof(MB_ROW_DEC) * pbi->decoding_thread_count);
        CHECK_MEM_ERROR(pbi->de_thread_data, vpx_malloc(sizeof(DECODETHREAD_DATA) * pbi->decoding_thread_count));
@@ -456,9 +457,12 @@ void vp8_decoder_create_threads(VP8D_COMP *pbi)
            pbi->de_thread_data[ithread].ptr1     = (void *)pbi;
            pbi->de_thread_data[ithread].ptr2     = (void *) &pbi->mb_row_di[ithread];

-            pthread_create(&pbi->h_decoding_thread[ithread], 0, vp8_thread_decoding_proc, (&pbi->de_thread_data[ithread]));
+            pthread_create(&pbi->h_decoding_thread[ithread], 0, thread_decoding_proc, (&pbi->de_thread_data[ithread]));
        }

+        for (ithread = 0; ithread < pbi->decoding_thread_count + 1; ithread++)
+            sem_init(&pbi->h_mb_counter[ithread], 0, 0);
+
        sem_init(&pbi->h_event_end_decoding, 0, 0);

        pbi->allocated_decoding_thread_count = pbi->decoding_thread_count;
@@ -473,22 +477,16 @@ void vp8mt_de_alloc_temp_buffers(VP8D_COMP *pbi, int mb_rows)

    if (pbi->b_multithreaded_rd)
    {
-        if (pbi->mt_current_mb_col)
-        {
            vpx_free(pbi->mt_current_mb_col);
            pbi->mt_current_mb_col = NULL ;
-        }

        /* Free above_row buffers. */
        if (pbi->mt_yabove_row)
        {
            for (i=0; i< mb_rows; i++)
            {
-                if (pbi->mt_yabove_row[i])
-                {
                    vpx_free(pbi->mt_yabove_row[i]);
                    pbi->mt_yabove_row[i] = NULL ;
-                }
            }
            vpx_free(pbi->mt_yabove_row);
            pbi->mt_yabove_row = NULL ;
@@ -498,11 +496,8 @@ void vp8mt_de_alloc_temp_buffers(VP8D_COMP *pbi, int mb_rows)
        {
            for (i=0; i< mb_rows; i++)
            {
-                if (pbi->mt_uabove_row[i])
-                {
                    vpx_free(pbi->mt_uabove_row[i]);
                    pbi->mt_uabove_row[i] = NULL ;
-                }
            }
            vpx_free(pbi->mt_uabove_row);
            pbi->mt_uabove_row = NULL ;
@@ -512,11 +507,8 @@ void vp8mt_de_alloc_temp_buffers(VP8D_COMP *pbi, int mb_rows)
        {
            for (i=0; i< mb_rows; i++)
            {
-                if (pbi->mt_vabove_row[i])
-                {
                    vpx_free(pbi->mt_vabove_row[i]);
                    pbi->mt_vabove_row[i] = NULL ;
-                }
            }
            vpx_free(pbi->mt_vabove_row);
            pbi->mt_vabove_row = NULL ;
@@ -527,11 +519,8 @@ void vp8mt_de_alloc_temp_buffers(VP8D_COMP *pbi, int mb_rows)
        {
            for (i=0; i< mb_rows; i++)
            {
-                if (pbi->mt_yleft_col[i])
-                {
                    vpx_free(pbi->mt_yleft_col[i]);
                    pbi->mt_yleft_col[i] = NULL ;
-                }
            }
            vpx_free(pbi->mt_yleft_col);
            pbi->mt_yleft_col = NULL ;
@@ -541,11 +530,8 @@ void vp8mt_de_alloc_temp_buffers(VP8D_COMP *pbi, int mb_rows)
        {
            for (i=0; i< mb_rows; i++)
            {
-                if (pbi->mt_uleft_col[i])
-                {
                    vpx_free(pbi->mt_uleft_col[i]);
                    pbi->mt_uleft_col[i] = NULL ;
-                }
            }
            vpx_free(pbi->mt_uleft_col);
            pbi->mt_uleft_col = NULL ;
@@ -555,11 +541,8 @@ void vp8mt_de_alloc_temp_buffers(VP8D_COMP *pbi, int mb_rows)
        {
            for (i=0; i< mb_rows; i++)
            {
-                if (pbi->mt_vleft_col[i])
-                {
                    vpx_free(pbi->mt_vleft_col[i]);
                    pbi->mt_vleft_col[i] = NULL ;
-                }
            }
            vpx_free(pbi->mt_vleft_col);
            pbi->mt_vleft_col = NULL ;
@@ -642,36 +625,32 @@ void vp8_decoder_remove_threads(VP8D_COMP *pbi)
            sem_destroy(&pbi->h_event_start_decoding[i]);
        }

+        for (i = 0; i < pbi->decoding_thread_count + 1; i++)
+            sem_destroy(&pbi->h_mb_counter[i]);
+
        sem_destroy(&pbi->h_event_end_decoding);

-        if (pbi->h_decoding_thread)
-        {
            vpx_free(pbi->h_decoding_thread);
            pbi->h_decoding_thread = NULL;
-        }

-        if (pbi->h_event_start_decoding)
-        {
            vpx_free(pbi->h_event_start_decoding);
            pbi->h_event_start_decoding = NULL;
-        }

-        if (pbi->mb_row_di)
+        if (pbi->h_mb_counter)
        {
+            vpx_free(pbi->h_mb_counter);
+            pbi->h_mb_counter = NULL;
+        }
            vpx_free(pbi->mb_row_di);
            pbi->mb_row_di = NULL ;
-        }

-        if (pbi->de_thread_data)
-        {
            vpx_free(pbi->de_thread_data);
            pbi->de_thread_data = NULL;
-        }
    }
 }


-void vp8mt_lpf_init( VP8D_COMP *pbi, int default_filt_lvl)
+static void lpf_init( VP8D_COMP *pbi, int default_filt_lvl)
 {
    VP8_COMMON *cm  = &pbi->common;
    MACROBLOCKD *mbd = &pbi->mb;
@@ -754,14 +733,17 @@ void vp8mt_decode_mb_rows( VP8D_COMP *pbi, MACROBLOCKD *xd)
            vpx_memset(pbi->mt_uleft_col[i], (unsigned char)129, 8);
            vpx_memset(pbi->mt_vleft_col[i], (unsigned char)129, 8);
        }
-        vp8mt_lpf_init(pbi, pc->filter_level);
+        lpf_init(pbi, pc->filter_level);
    }

-    vp8_setup_decoding_thread_data(pbi, xd, pbi->mb_row_di, pbi->decoding_thread_count);
+    setup_decoding_thread_data(pbi, xd, pbi->mb_row_di, pbi->decoding_thread_count);
+
+

    for (i = 0; i < pbi->decoding_thread_count; i++)
        sem_post(&pbi->h_event_start_decoding[i]);

+
    for (mb_row = 0; mb_row < pc->mb_rows; mb_row += (pbi->decoding_thread_count + 1))
    {

@@ -794,6 +776,7 @@ void vp8mt_decode_mb_rows( VP8D_COMP *pbi, MACROBLOCKD *xd)

            for (mb_col = 0; mb_col < pc->mb_cols; mb_col++)
            {
+/*
                if ( mb_row > 0 && (mb_col & (nsync-1)) == 0){
                    while (mb_col > (*last_row_current_mb_col - nsync) && *last_row_current_mb_col != pc->mb_cols - 1)
                    {
@@ -801,6 +784,9 @@ void vp8mt_decode_mb_rows( VP8D_COMP *pbi, MACROBLOCKD *xd)
                        thread_sleep(0);
                    }
                }
+*/
+                if(mb_row > 0)
+                    sem_wait(&pbi->h_mb_counter[pbi->decoding_thread_count]);

                if (xd->mode_info_context->mbmi.mode == SPLITMV || xd->mode_info_context->mbmi.mode == B_PRED)
                {
@@ -811,18 +797,6 @@ void vp8mt_decode_mb_rows( VP8D_COMP *pbi, MACROBLOCKD *xd)
                    }
                }

-                if(pbi->common.filter_level)
-                {
-                    /* update loopfilter info */
-                    Segment = (alt_flt_enabled) ? xd->mode_info_context->mbmi.segment_id : 0;
-                    filter_level = pbi->mt_baseline_filter_level[Segment];
-                    /* Distance of Mb to the various image edges.
-                     * These are specified to 8th pel as they are always compared to values that are in 1/8th pel units
-                     * Apply any context driven MB level adjustment
-                     */
-                    filter_level = vp8_adjust_mb_lf_value(xd, filter_level);
-                }
-
                /* Distance of Mb to the various image edges.
                 * These are specified to 8th pel as they are always compared to values that are in 1/8th pel units
                 */
@@ -854,7 +828,7 @@ void vp8mt_decode_mb_rows( VP8D_COMP *pbi, MACROBLOCKD *xd)
                }

                vp8_build_uvmvs(xd, pc->full_pixel);
-                vp8mt_decode_macroblock(pbi, xd, mb_row, mb_col);
+                decode_macroblock(pbi, xd, mb_row, mb_col);

                /* check if the boolean decoder has suffered an error */
                xd->corrupted |= vp8dx_bool_error(xd->current_bc);
@@ -886,6 +860,15 @@ void vp8mt_decode_mb_rows( VP8D_COMP *pbi, MACROBLOCKD *xd)
                        }
                    }

+                    /* update loopfilter info */
+                    Segment = (alt_flt_enabled) ? xd->mode_info_context->mbmi.segment_id : 0;
+                    filter_level = pbi->mt_baseline_filter_level[Segment];
+                    /* Distance of Mb to the various image edges.
+                     * These are specified to 8th pel as they are always compared to values that are in 1/8th pel units
+                     * Apply any context driven MB level adjustment
+                     */
+                    filter_level = vp8_adjust_mb_lf_value(xd, filter_level);
+
                    /* loopfilter on this macroblock. */
                    if (filter_level)
                    {
@@ -912,6 +895,10 @@ void vp8mt_decode_mb_rows( VP8D_COMP *pbi, MACROBLOCKD *xd)
                xd->above_context++;

                pbi->mt_current_mb_col[mb_row] = mb_col;
+
+                /* macroblock counter */
+                if (mb_row != pbi->common.mb_rows-1)
+                    sem_post(&pbi->h_mb_counter[0]);
            }

            /* adjust to the next row of mbs */
--- a/vp8/decoder/treereader.h
+++ b/vp8/decoder/treereader.h
@@ -12,7 +12,7 @@
 #ifndef tree_reader_h
 #define tree_reader_h 1

-#include "treecoder.h"
+#include "vp8/common/treecoder.h"

 #include "dboolhuff.h"

--- a/vp8/decoder/x86/idct_blk_mmx.c
+++ b/vp8/decoder/x86/idct_blk_mmx.c
@@ -9,8 +9,8 @@
 */

 #include "vpx_ports/config.h"
-#include "idct.h"
-#include "dequantize.h"
+#include "vp8/common/idct.h"
+#include "vp8/decoder/dequantize.h"

 void vp8_dequant_dc_idct_add_y_block_mmx
            (short *q, short *dq, unsigned char *pre,
--- a/vp8/decoder/x86/idct_blk_sse2.c
+++ b/vp8/decoder/x86/idct_blk_sse2.c
@@ -9,8 +9,8 @@
 */

 #include "vpx_ports/config.h"
-#include "idct.h"
-#include "dequantize.h"
+#include "vp8/common/idct.h"
+#include "vp8/decoder/dequantize.h"

 void idct_dequant_dc_0_2x_sse2
            (short *q, short *dq, unsigned char *pre,
--- a/vp8/decoder/x86/onyxdxv.c
+++ b/vp8/decoder/x86/onyxdxv.c
--- a/vp8/decoder/x86/x86_dsystemdependent.c
+++ b/vp8/decoder/x86/x86_dsystemdependent.c
@@ -11,13 +11,13 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/x86.h"
-#include "onyxd_int.h"
+#include "vp8/decoder/onyxd_int.h"


 #if HAVE_MMX
 void vp8_dequantize_b_impl_mmx(short *sq, short *dq, short *q);

-void vp8_dequantize_b_mmx(BLOCKD *d)
+static void dequantize_b_mmx(BLOCKD *d)
 {
    short *sq = (short *) d->qcoeff;
    short *dq = (short *) d->dqcoeff;
@@ -41,7 +41,7 @@ void vp8_arch_x86_decode_init(VP8D_COMP *pbi)
 #if HAVE_MMX
    if (flags & HAS_MMX)
    {
-        pbi->dequant.block               = vp8_dequantize_b_mmx;
+        pbi->dequant.block               = dequantize_b_mmx;
        pbi->dequant.idct_add            = vp8_dequant_idct_add_mmx;
        pbi->dequant.dc_idct_add         = vp8_dequant_dc_idct_add_mmx;
        pbi->dequant.dc_idct_add_y_block = vp8_dequant_dc_idct_add_y_block_mmx;
--- a/vp8/encoder/arm/arm_csystemdependent.c
+++ b/vp8/encoder/arm/arm_csystemdependent.c
@@ -11,8 +11,8 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/arm.h"
-#include "variance.h"
-#include "onyx_int.h"
+#include "vp8/encoder/variance.h"
+#include "vp8/encoder/onyx_int.h"

 extern void (*vp8_yv12_copy_partial_frame_ptr)(YV12_BUFFER_CONFIG *src_ybc, YV12_BUFFER_CONFIG *dst_ybc, int Fraction);
 extern void vp8_yv12_copy_partial_frame(YV12_BUFFER_CONFIG *src_ybc, YV12_BUFFER_CONFIG *dst_ybc, int Fraction);
@@ -29,26 +29,29 @@ void vp8_arch_arm_encoder_init(VP8_COMP *cpi)
 #if HAVE_ARMV6
    if (has_media)
    {
-        /*cpi->rtcd.variance.sad16x16              = vp8_sad16x16_c;
-        cpi->rtcd.variance.sad16x8               = vp8_sad16x8_c;
+        cpi->rtcd.variance.sad16x16              = vp8_sad16x16_armv6;
+        /*cpi->rtcd.variance.sad16x8               = vp8_sad16x8_c;
        cpi->rtcd.variance.sad8x16               = vp8_sad8x16_c;
        cpi->rtcd.variance.sad8x8                = vp8_sad8x8_c;
        cpi->rtcd.variance.sad4x4                = vp8_sad4x4_c;*/

-        /*cpi->rtcd.variance.var4x4                = vp8_variance4x4_c;
-        cpi->rtcd.variance.var8x8                = vp8_variance8x8_c;
-        cpi->rtcd.variance.var8x16               = vp8_variance8x16_c;
+        /*cpi->rtcd.variance.var4x4                = vp8_variance4x4_c;*/
+        cpi->rtcd.variance.var8x8                = vp8_variance8x8_armv6;
+        /*cpi->rtcd.variance.var8x16               = vp8_variance8x16_c;
        cpi->rtcd.variance.var16x8               = vp8_variance16x8_c;*/
        cpi->rtcd.variance.var16x16              = vp8_variance16x16_armv6;

-        /*cpi->rtcd.variance.subpixvar4x4          = vp8_sub_pixel_variance4x4_c;
-        cpi->rtcd.variance.subpixvar8x8          = vp8_sub_pixel_variance8x8_c;
-        cpi->rtcd.variance.subpixvar8x16         = vp8_sub_pixel_variance8x16_c;
+        /*cpi->rtcd.variance.subpixvar4x4          = vp8_sub_pixel_variance4x4_c;*/
+        cpi->rtcd.variance.subpixvar8x8          = vp8_sub_pixel_variance8x8_armv6;
+        /*cpi->rtcd.variance.subpixvar8x16         = vp8_sub_pixel_variance8x16_c;
        cpi->rtcd.variance.subpixvar16x8         = vp8_sub_pixel_variance16x8_c;*/
        cpi->rtcd.variance.subpixvar16x16        = vp8_sub_pixel_variance16x16_armv6;
+        cpi->rtcd.variance.halfpixvar16x16_h     = vp8_variance_halfpixvar16x16_h_armv6;
+        cpi->rtcd.variance.halfpixvar16x16_v     = vp8_variance_halfpixvar16x16_v_armv6;
+        cpi->rtcd.variance.halfpixvar16x16_hv    = vp8_variance_halfpixvar16x16_hv_armv6;

-        /*cpi->rtcd.variance.mse16x16              = vp8_mse16x16_c;
-        cpi->rtcd.variance.getmbss               = vp8_get_mb_ss_c;*/
+        cpi->rtcd.variance.mse16x16              = vp8_mse16x16_armv6;
+        /*cpi->rtcd.variance.getmbss               = vp8_get_mb_ss_c;*/

        /*cpi->rtcd.variance.get16x16prederror     = vp8_get16x16pred_error_c;
        cpi->rtcd.variance.get8x8var             = vp8_get8x8var_c;
@@ -56,9 +59,9 @@ void vp8_arch_arm_encoder_init(VP8_COMP *cpi)
        cpi->rtcd.variance.get4x4sse_cs          = vp8_get4x4sse_cs_c;*/

        /*cpi->rtcd.fdct.short4x4                  = vp8_short_fdct4x4_c;
-        cpi->rtcd.fdct.short8x4                  = vp8_short_fdct8x4_c;
-        cpi->rtcd.fdct.fast4x4                   = vp8_fast_fdct4x4_c;
-        cpi->rtcd.fdct.fast8x4                   = vp8_fast_fdct8x4_c;*/
+        cpi->rtcd.fdct.short8x4                  = vp8_short_fdct8x4_c;*/
+        cpi->rtcd.fdct.fast4x4                   = vp8_fast_fdct4x4_armv6;
+        cpi->rtcd.fdct.fast8x4                   = vp8_fast_fdct8x4_armv6;
        cpi->rtcd.fdct.walsh_short4x4            = vp8_short_walsh4x4_armv6;

        /*cpi->rtcd.encodemb.berr                  = vp8_block_error_c;
@@ -68,8 +71,8 @@ void vp8_arch_arm_encoder_init(VP8_COMP *cpi)
        cpi->rtcd.encodemb.submby                = vp8_subtract_mby_c;
        cpi->rtcd.encodemb.submbuv               = vp8_subtract_mbuv_c;*/

-        /*cpi->rtcd.quantize.quantb                = vp8_regular_quantize_b;
-        cpi->rtcd.quantize.fastquantb            = vp8_fast_quantize_b_c;*/
+        /*cpi->rtcd.quantize.quantb                = vp8_regular_quantize_b;*/
+        cpi->rtcd.quantize.fastquantb            = vp8_fast_quantize_b_armv6;
    }
 #endif

--- a/vp8/encoder/arm/armv6/vp8_fast_fdct4x4_armv6.asm
+++ b/vp8/encoder/arm/armv6/vp8_fast_fdct4x4_armv6.asm
@@ -0,0 +1,262 @@
+;
+;  Copyright (c) 2011 The WebM project authors. All Rights Reserved.
+;
+;  Use of this source code is governed by a BSD-style license
+;  that can be found in the LICENSE file in the root of the source
+;  tree. An additional intellectual property rights grant can be found
+;  in the file PATENTS.  All contributing project authors may
+;  be found in the AUTHORS file in the root of the source tree.
+;
+
+    EXPORT |vp8_fast_fdct4x4_armv6|
+
+    ARM
+    REQUIRE8
+    PRESERVE8
+
+    AREA    |.text|, CODE, READONLY
+; void vp8_short_fdct4x4_c(short *input, short *output, int pitch)
+|vp8_fast_fdct4x4_armv6| PROC
+
+    stmfd       sp!, {r4 - r12, lr}
+
+    ; PART 1
+
+    ; coeffs 0-3
+    ldrd        r4, r5, [r0]        ; [i1 | i0] [i3 | i2]
+
+    ldr         r10, c7500
+    ldr         r11, c14500
+    ldr         r12, c0x22a453a0    ; [2217*4 | 5352*4]
+    ldr         lr, c0x00080008
+    ror         r5, r5, #16         ; [i2 | i3]
+
+    qadd16      r6, r4, r5          ; [i1+i2 | i0+i3] = [b1 | a1] without shift
+    qsub16      r7, r4, r5          ; [i1-i2 | i0-i3] = [c1 | d1] without shift
+
+    add         r0, r0, r2          ; update input pointer
+
+    qadd16      r7, r7, r7          ; 2*[c1|d1] --> we can use smlad and smlsd
+                                    ; with 2217*4 and 5352*4 without losing the
+                                    ; sign bit (overflow)
+
+    smuad       r4, r6, lr          ; o0 = (i1+i2)*8 + (i0+i3)*8
+    smusd       r5, r6, lr          ; o2 = (i1+i2)*8 - (i0+i3)*8
+
+    smlad       r6, r7, r12, r11    ; o1 = (c1 * 2217 + d1 * 5352 +  14500)
+    smlsdx      r7, r7, r12, r10    ; o3 = (d1 * 2217 - c1 * 5352 +   7500)
+
+    ldrd        r8, r9, [r0]        ; [i5 | i4] [i7 | i6]
+
+    pkhbt       r3, r4, r6, lsl #4  ; [o1 | o0], keep in register for PART 2
+    pkhbt       r6, r5, r7, lsl #4  ; [o3 | o2]
+
+    str         r6, [r1, #4]
+
+    ; coeffs 4-7
+    ror         r9, r9, #16         ; [i6 | i7]
+
+    qadd16      r6, r8, r9          ; [i5+i6 | i4+i7] = [b1 | a1] without shift
+    qsub16      r7, r8, r9          ; [i5-i6 | i4-i7] = [c1 | d1] without shift
+
+    add         r0, r0, r2          ; update input pointer
+
+    qadd16      r7, r7, r7          ; 2x[c1|d1] --> we can use smlad and smlsd
+                                    ; with 2217*4 and 5352*4 without losing the
+                                    ; sign bit (overflow)
+
+    smuad       r9, r6, lr          ; o4 = (i5+i6)*8 + (i4+i7)*8
+    smusd       r8, r6, lr          ; o6 = (i5+i6)*8 - (i4+i7)*8
+
+    smlad       r6, r7, r12, r11    ; o5 = (c1 * 2217 + d1 * 5352 +  14500)
+    smlsdx      r7, r7, r12, r10    ; o7 = (d1 * 2217 - c1 * 5352 +   7500)
+
+    ldrd        r4, r5, [r0]        ; [i9 | i8] [i11 | i10]
+
+    pkhbt       r9, r9, r6, lsl #4  ; [o5 | o4], keep in register for PART 2
+    pkhbt       r6, r8, r7, lsl #4  ; [o7 | o6]
+
+    str         r6, [r1, #12]
+
+    ; coeffs 8-11
+    ror         r5, r5, #16         ; [i10 | i11]
+
+    qadd16      r6, r4, r5          ; [i9+i10 | i8+i11]=[b1 | a1] without shift
+    qsub16      r7, r4, r5          ; [i9-i10 | i8-i11]=[c1 | d1] without shift
+
+    add         r0, r0, r2          ; update input pointer
+
+    qadd16      r7, r7, r7          ; 2x[c1|d1] --> we can use smlad and smlsd
+                                    ; with 2217*4 and 5352*4 without losing the
+                                    ; sign bit (overflow)
+
+    smuad       r2, r6, lr          ; o8 = (i9+i10)*8 + (i8+i11)*8
+    smusd       r8, r6, lr          ; o10 = (i9+i10)*8 - (i8+i11)*8
+
+    smlad       r6, r7, r12, r11    ; o9 = (c1 * 2217 + d1 * 5352 +  14500)
+    smlsdx      r7, r7, r12, r10    ; o11 = (d1 * 2217 - c1 * 5352 +   7500)
+
+    ldrd        r4, r5, [r0]        ; [i13 | i12] [i15 | i14]
+
+    pkhbt       r2, r2, r6, lsl #4  ; [o9 | o8], keep in register for PART 2
+    pkhbt       r6, r8, r7, lsl #4  ; [o11 | o10]
+
+    str         r6, [r1, #20]
+
+    ; coeffs 12-15
+    ror         r5, r5, #16         ; [i14 | i15]
+
+    qadd16      r6, r4, r5          ; [i13+i14 | i12+i15]=[b1|a1] without shift
+    qsub16      r7, r4, r5          ; [i13-i14 | i12-i15]=[c1|d1] without shift
+
+    qadd16      r7, r7, r7          ; 2x[c1|d1] --> we can use smlad and smlsd
+                                    ; with 2217*4 and 5352*4 without losing the
+                                    ; sign bit (overflow)
+
+    smuad       r4, r6, lr          ; o12 = (i13+i14)*8 + (i12+i15)*8
+    smusd       r5, r6, lr          ; o14 = (i13+i14)*8 - (i12+i15)*8
+
+    smlad       r6, r7, r12, r11    ; o13 = (c1 * 2217 + d1 * 5352 +  14500)
+    smlsdx      r7, r7, r12, r10    ; o15 = (d1 * 2217 - c1 * 5352 +   7500)
+
+    pkhbt       r0, r4, r6, lsl #4  ; [o13 | o12], keep in register for PART 2
+    pkhbt       r6, r5, r7, lsl #4  ; [o15 | o14]
+
+    str         r6, [r1, #28]
+
+
+    ; PART 2 -------------------------------------------------
+    ldr         r11, c12000
+    ldr         r10, c51000
+    ldr         lr, c0x00070007
+
+    qadd16      r4, r3, r0          ; a1 = [i1+i13 | i0+i12]
+    qadd16      r5, r9, r2          ; b1 = [i5+i9  |  i4+i8]
+    qsub16      r6, r9, r2          ; c1 = [i5-i9  |  i4-i8]
+    qsub16      r7, r3, r0          ; d1 = [i1-i13 | i0-i12]
+
+    qadd16      r4, r4, lr          ; a1 + 7
+
+    add         r0, r11, #0x10000   ; add (d!=0)
+
+    qadd16      r2, r4, r5          ; a1 + b1 + 7
+    qsub16      r3, r4, r5          ; a1 - b1 + 7
+
+    ldr         r12, c0x08a914e8    ; [2217 | 5352]
+
+    lsl         r8, r2, #16         ; prepare bottom halfword for scaling
+    asr         r2, r2, #4          ; scale top halfword
+    lsl         r9, r3, #16         ; prepare bottom halfword for scaling
+    asr         r3, r3, #4          ; scale top halfword
+    pkhtb       r4, r2, r8, asr #20 ; pack and scale bottom halfword
+    pkhtb       r5, r3, r9, asr #20 ; pack and scale bottom halfword
+
+    smulbt      r2, r6, r12         ; [ ------ | c1*2217]
+    str         r4, [r1, #0]        ; [     o1 |      o0]
+    smultt      r3, r6, r12         ; [c1*2217 | ------ ]
+    str         r5, [r1, #16]       ; [     o9 |      o8]
+
+    smlabb      r8, r7, r12, r2     ; [ ------ | d1*5352]
+    smlatb      r9, r7, r12, r3     ; [d1*5352 | ------ ]
+
+    smulbb      r2, r6, r12         ; [ ------ | c1*5352]
+    smultb      r3, r6, r12         ; [c1*5352 | ------ ]
+
+    lsls        r6, r7, #16         ; d1 != 0 ?
+    addeq       r8, r8, r11         ; c1_b*2217+d1_b*5352+12000 + (d==0)
+    addne       r8, r8, r0          ; c1_b*2217+d1_b*5352+12000 + (d!=0)
+    asrs        r6, r7, #16
+    addeq       r9, r9, r11         ; c1_t*2217+d1_t*5352+12000 + (d==0)
+    addne       r9, r9, r0          ; c1_t*2217+d1_t*5352+12000 + (d!=0)
+
+    smlabt      r4, r7, r12, r10    ; [ ------ | d1*2217] + 51000
+    smlatt      r5, r7, r12, r10    ; [d1*2217 | ------ ] + 51000
+
+    pkhtb       r9, r9, r8, asr #16
+
+    sub         r4, r4, r2
+    sub         r5, r5, r3
+
+    ldr         r3, [r1, #4]        ; [i3 | i2]
+
+    pkhtb       r5, r5, r4, asr #16 ; [o13|o12]
+
+    str         r9, [r1, #8]        ; [o5 | 04]
+
+    ldr         r9, [r1, #12]       ; [i7 | i6]
+    ldr         r8, [r1, #28]       ; [i15|i14]
+    ldr         r2, [r1, #20]       ; [i11|i10]
+    str         r5, [r1, #24]       ; [o13|o12]
+
+    qadd16      r4, r3, r8          ; a1 = [i3+i15 | i2+i14]
+    qadd16      r5, r9, r2          ; b1 = [i7+i11 | i6+i10]
+
+    qadd16      r4, r4, lr          ; a1 + 7
+
+    qsub16      r6, r9, r2          ; c1 = [i7-i11 | i6-i10]
+    qadd16      r2, r4, r5          ; a1 + b1 + 7
+    qsub16      r7, r3, r8          ; d1 = [i3-i15 | i2-i14]
+    qsub16      r3, r4, r5          ; a1 - b1 + 7
+
+    lsl         r8, r2, #16         ; prepare bottom halfword for scaling
+    asr         r2, r2, #4          ; scale top halfword
+    lsl         r9, r3, #16         ; prepare bottom halfword for scaling
+    asr         r3, r3, #4          ; scale top halfword
+    pkhtb       r4, r2, r8, asr #20 ; pack and scale bottom halfword
+    pkhtb       r5, r3, r9, asr #20 ; pack and scale bottom halfword
+
+    smulbt      r2, r6, r12         ; [ ------ | c1*2217]
+    str         r4, [r1, #4]        ; [     o3 |      o2]
+    smultt      r3, r6, r12         ; [c1*2217 | ------ ]
+    str         r5, [r1, #20]       ; [    o11 |     o10]
+
+    smlabb      r8, r7, r12, r2     ; [ ------ | d1*5352]
+    smlatb      r9, r7, r12, r3     ; [d1*5352 | ------ ]
+
+    smulbb      r2, r6, r12         ; [ ------ | c1*5352]
+    smultb      r3, r6, r12         ; [c1*5352 | ------ ]
+
+    lsls        r6, r7, #16         ; d1 != 0 ?
+    addeq       r8, r8, r11         ; c1_b*2217+d1_b*5352+12000 + (d==0)
+    addne       r8, r8, r0          ; c1_b*2217+d1_b*5352+12000 + (d!=0)
+
+    asrs        r6, r7, #16
+    addeq       r9, r9, r11         ; c1_t*2217+d1_t*5352+12000 + (d==0)
+    addne       r9, r9, r0          ; c1_t*2217+d1_t*5352+12000 + (d!=0)
+
+    smlabt      r4, r7, r12, r10    ; [ ------ | d1*2217] + 51000
+    smlatt      r5, r7, r12, r10    ; [d1*2217 | ------ ] + 51000
+
+    pkhtb       r9, r9, r8, asr #16
+
+    sub         r4, r4, r2
+    sub         r5, r5, r3
+
+    str         r9, [r1, #12]       ; [o7 | o6]
+    pkhtb       r5, r5, r4, asr #16 ; [o15|o14]
+
+    str         r5, [r1, #28]       ; [o15|o14]
+
+    ldmfd       sp!, {r4 - r12, pc}
+
+    ENDP
+
+; Used constants
+c7500
+    DCD     7500
+c14500
+    DCD     14500
+c0x22a453a0
+    DCD     0x22a453a0
+c0x00080008
+    DCD     0x00080008
+c12000
+    DCD     12000
+c51000
+    DCD     51000
+c0x00070007
+    DCD     0x00070007
+c0x08a914e8
+    DCD     0x08a914e8
+
+    END
--- a/vp8/encoder/arm/armv6/vp8_fast_quantize_b_armv6.asm
+++ b/vp8/encoder/arm/armv6/vp8_fast_quantize_b_armv6.asm
@@ -0,0 +1,224 @@
+;
+;  Copyright (c) 2011 The WebM project authors. All Rights Reserved.
+;
+;  Use of this source code is governed by a BSD-style license
+;  that can be found in the LICENSE file in the root of the source
+;  tree. An additional intellectual property rights grant can be found
+;  in the file PATENTS.  All contributing project authors may
+;  be found in the AUTHORS file in the root of the source tree.
+;
+
+
+    EXPORT  |vp8_fast_quantize_b_armv6|
+
+    INCLUDE asm_enc_offsets.asm
+
+    ARM
+    REQUIRE8
+    PRESERVE8
+
+    AREA ||.text||, CODE, READONLY, ALIGN=2
+
+; r0    BLOCK *b
+; r1    BLOCKD *d
+|vp8_fast_quantize_b_armv6| PROC
+    stmfd   sp!, {r1, r4-r11, lr}
+
+    ldr     r3, [r0, #vp8_block_coeff]      ; coeff
+    ldr     r4, [r0, #vp8_block_quant_fast] ; quant_fast
+    ldr     r5, [r0, #vp8_block_round]      ; round
+    ldr     r6, [r1, #vp8_blockd_qcoeff]    ; qcoeff
+    ldr     r7, [r1, #vp8_blockd_dqcoeff]   ; dqcoeff
+    ldr     r8, [r1, #vp8_blockd_dequant]   ; dequant
+
+    ldr     r2, loop_count          ; loop_count=0x1000000. 'lsls' instruction
+                                    ; is used to update the counter so that
+                                    ; it can be used to mark nonzero
+                                    ; quantized coefficient pairs.
+
+    mov     r1, #0                  ; flags for quantized coeffs
+
+    ; PART 1: quantization and dequantization loop
+loop
+    ldr     r9, [r3], #4            ; [z1 | z0]
+    ldr     r10, [r5], #4           ; [r1 | r0]
+    ldr     r11, [r4], #4           ; [q1 | q0]
+
+    ssat16  lr, #1, r9              ; [sz1 | sz0]
+    eor     r9, r9, lr              ; [z1 ^ sz1 | z0 ^ sz0]
+    ssub16  r9, r9, lr              ; x = (z ^ sz) - sz
+    sadd16  r9, r9, r10             ; [x1+r1 | x0+r0]
+
+    ldr     r12, [r3], #4           ; [z3 | z2]
+
+    smulbb  r0, r9, r11             ; [(x0+r0)*q0]
+    smultt  r9, r9, r11             ; [(x1+r1)*q1]
+
+    ldr     r10, [r5], #4           ; [r3 | r2]
+
+    ssat16  r11, #1, r12            ; [sz3 | sz2]
+    eor     r12, r12, r11           ; [z3 ^ sz3 | z2 ^ sz2]
+    pkhtb   r0, r9, r0, asr #16     ; [y1 | y0]
+    ldr     r9, [r4], #4            ; [q3 | q2]
+    ssub16  r12, r12, r11           ; x = (z ^ sz) - sz
+
+    sadd16  r12, r12, r10           ; [x3+r3 | x2+r2]
+
+    eor     r0, r0, lr              ; [(y1 ^ sz1) | (y0 ^ sz0)]
+
+    smulbb  r10, r12, r9            ; [(x2+r2)*q2]
+    smultt  r12, r12, r9            ; [(x3+r3)*q3]
+
+    ssub16  r0, r0, lr              ; x = (y ^ sz) - sz
+
+    cmp     r0, #0                  ; check if zero
+    orrne   r1, r1, r2, lsr #24     ; add flag for nonzero coeffs
+
+    str     r0, [r6], #4            ; *qcoeff++ = x
+    ldr     r9, [r8], #4            ; [dq1 | dq0]
+
+    pkhtb   r10, r12, r10, asr #16  ; [y3 | y2]
+    eor     r10, r10, r11           ; [(y3 ^ sz3) | (y2 ^ sz2)]
+    ssub16  r10, r10, r11           ; x = (y ^ sz) - sz
+
+    cmp     r10, #0                 ; check if zero
+    orrne   r1, r1, r2, lsr #23     ; add flag for nonzero coeffs
+
+    str     r10, [r6], #4           ; *qcoeff++ = x
+    ldr     r11, [r8], #4           ; [dq3 | dq2]
+
+    smulbb  r12, r0, r9             ; [x0*dq0]
+    smultt  r0, r0, r9              ; [x1*dq1]
+
+    smulbb  r9, r10, r11            ; [x2*dq2]
+    smultt  r10, r10, r11           ; [x3*dq3]
+
+    lsls    r2, r2, #2              ; update loop counter
+    strh    r12, [r7, #0]           ; dqcoeff[0] = [x0*dq0]
+    strh    r0, [r7, #2]            ; dqcoeff[1] = [x1*dq1]
+    strh    r9, [r7, #4]            ; dqcoeff[2] = [x2*dq2]
+    strh    r10, [r7, #6]           ; dqcoeff[3] = [x3*dq3]
+    add     r7, r7, #8              ; dqcoeff += 8
+    bne     loop
+
+    ; PART 2: check position for eob...
+    mov     lr, #0                  ; init eob
+    cmp     r1, #0                  ; coeffs after quantization?
+    ldr     r11, [sp, #0]           ; restore BLOCKD pointer
+    beq     end                     ; skip eob calculations if all zero
+
+    ldr     r0, [r11, #vp8_blockd_qcoeff]
+
+    ; check shortcut for nonzero qcoeffs
+    tst    r1, #0x80
+    bne    quant_coeff_15_14
+    tst    r1, #0x20
+    bne    quant_coeff_13_11
+    tst    r1, #0x8
+    bne    quant_coeff_12_7
+    tst    r1, #0x40
+    bne    quant_coeff_10_9
+    tst    r1, #0x10
+    bne    quant_coeff_8_3
+    tst    r1, #0x2
+    bne    quant_coeff_6_5
+    tst    r1, #0x4
+    bne    quant_coeff_4_2
+    b      quant_coeff_1_0
+
+quant_coeff_15_14
+    ldrh    r2, [r0, #30]       ; rc=15, i=15
+    mov     lr, #16
+    cmp     r2, #0
+    bne     end
+
+    ldrh    r3, [r0, #28]       ; rc=14, i=14
+    mov     lr, #15
+    cmp     r3, #0
+    bne     end
+
+quant_coeff_13_11
+    ldrh    r2, [r0, #22]       ; rc=11, i=13
+    mov     lr, #14
+    cmp     r2, #0
+    bne     end
+
+quant_coeff_12_7
+    ldrh    r3, [r0, #14]       ; rc=7,  i=12
+    mov     lr, #13
+    cmp     r3, #0
+    bne     end
+
+    ldrh    r2, [r0, #20]       ; rc=10, i=11
+    mov     lr, #12
+    cmp     r2, #0
+    bne     end
+
+quant_coeff_10_9
+    ldrh    r3, [r0, #26]       ; rc=13, i=10
+    mov     lr, #11
+    cmp     r3, #0
+    bne     end
+
+    ldrh    r2, [r0, #24]       ; rc=12, i=9
+    mov     lr, #10
+    cmp     r2, #0
+    bne     end
+
+quant_coeff_8_3
+    ldrh    r3, [r0, #18]       ; rc=9,  i=8
+    mov     lr, #9
+    cmp     r3, #0
+    bne     end
+
+    ldrh    r2, [r0, #12]       ; rc=6,  i=7
+    mov     lr, #8
+    cmp     r2, #0
+    bne     end
+
+quant_coeff_6_5
+    ldrh    r3, [r0, #6]        ; rc=3,  i=6
+    mov     lr, #7
+    cmp     r3, #0
+    bne     end
+
+    ldrh    r2, [r0, #4]        ; rc=2,  i=5
+    mov     lr, #6
+    cmp     r2, #0
+    bne     end
+
+quant_coeff_4_2
+    ldrh    r3, [r0, #10]       ; rc=5,  i=4
+    mov     lr, #5
+    cmp     r3, #0
+    bne     end
+
+    ldrh    r2, [r0, #16]       ; rc=8,  i=3
+    mov     lr, #4
+    cmp     r2, #0
+    bne     end
+
+    ldrh    r3, [r0, #8]        ; rc=4,  i=2
+    mov     lr, #3
+    cmp     r3, #0
+    bne     end
+
+quant_coeff_1_0
+    ldrh    r2, [r0, #2]        ; rc=1,  i=1
+    mov     lr, #2
+    cmp     r2, #0
+    bne     end
+
+    mov     lr, #1              ; rc=0,  i=0
+
+end
+    str     lr, [r11, #vp8_blockd_eob]
+    ldmfd   sp!, {r1, r4-r11, pc}
+
+    ENDP
+
+loop_count
+    DCD     0x1000000
+
+    END
+
--- a/vp8/encoder/arm/armv6/vp8_mse16x16_armv6.asm
+++ b/vp8/encoder/arm/armv6/vp8_mse16x16_armv6.asm
@@ -0,0 +1,133 @@
+;
+;  Copyright (c) 2011 The WebM project authors. All Rights Reserved.
+;
+;  Use of this source code is governed by a BSD-style license
+;  that can be found in the LICENSE file in the root of the source
+;  tree. An additional intellectual property rights grant can be found
+;  in the file PATENTS.  All contributing project authors may
+;  be found in the AUTHORS file in the root of the source tree.
+;
+
+
+    EXPORT  |vp8_mse16x16_armv6|
+
+    ARM
+
+    AREA ||.text||, CODE, READONLY, ALIGN=2
+
+; r0    unsigned char *src_ptr
+; r1    int source_stride
+; r2    unsigned char *ref_ptr
+; r3    int  recon_stride
+; stack unsigned int *sse
+;
+;note: Based on vp8_variance16x16_armv6. In this function, sum is never used.
+;      So, we can remove this part of calculation.
+
+|vp8_mse16x16_armv6| PROC
+
+    push    {r4-r9, lr}
+    mov     r12, #16            ; set loop counter to 16 (=block height)
+
+    mov     r4, #0              ; initialize sse = 0
+
+loop
+    ; 1st 4 pixels
+    ldr     r5, [r0, #0x0]      ; load 4 src pixels
+    ldr     r6, [r2, #0x0]      ; load 4 ref pixels
+
+    mov     lr, #0              ; constant zero
+
+    usub8   r8, r5, r6          ; calculate difference
+    sel     r7, r8, lr          ; select bytes with positive difference
+    usub8   r9, r6, r5          ; calculate difference with reversed operands
+    sel     r8, r9, lr          ; select bytes with negative difference
+
+    ; calculate partial sums
+    usad8   r5, r7, lr          ; calculate sum of positive differences
+    usad8   r6, r8, lr          ; calculate sum of negative differences
+    orr     r8, r8, r7          ; differences of all 4 pixels
+
+    ldr     r5, [r0, #0x4]      ; load 4 src pixels
+
+    ; calculate sse
+    uxtb16  r6, r8              ; byte (two pixels) to halfwords
+    uxtb16  r7, r8, ror #8      ; another two pixels to halfwords
+    smlad   r4, r6, r6, r4      ; dual signed multiply, add and accumulate (1)
+
+    ; 2nd 4 pixels
+    ldr     r6, [r2, #0x4]      ; load 4 ref pixels
+    smlad   r4, r7, r7, r4      ; dual signed multiply, add and accumulate (2)
+
+    usub8   r8, r5, r6          ; calculate difference
+    sel     r7, r8, lr          ; select bytes with positive difference
+    usub8   r9, r6, r5          ; calculate difference with reversed operands
+    sel     r8, r9, lr          ; select bytes with negative difference
+
+    ; calculate partial sums
+    usad8   r5, r7, lr          ; calculate sum of positive differences
+    usad8   r6, r8, lr          ; calculate sum of negative differences
+    orr     r8, r8, r7          ; differences of all 4 pixels
+    ldr     r5, [r0, #0x8]      ; load 4 src pixels
+    ; calculate sse
+    uxtb16  r6, r8              ; byte (two pixels) to halfwords
+    uxtb16  r7, r8, ror #8      ; another two pixels to halfwords
+    smlad   r4, r6, r6, r4      ; dual signed multiply, add and accumulate (1)
+
+    ; 3rd 4 pixels
+    ldr     r6, [r2, #0x8]      ; load 4 ref pixels
+    smlad   r4, r7, r7, r4      ; dual signed multiply, add and accumulate (2)
+
+    usub8   r8, r5, r6          ; calculate difference
+    sel     r7, r8, lr          ; select bytes with positive difference
+    usub8   r9, r6, r5          ; calculate difference with reversed operands
+    sel     r8, r9, lr          ; select bytes with negative difference
+
+    ; calculate partial sums
+    usad8   r5, r7, lr          ; calculate sum of positive differences
+    usad8   r6, r8, lr          ; calculate sum of negative differences
+    orr     r8, r8, r7          ; differences of all 4 pixels
+
+    ldr     r5, [r0, #0xc]      ; load 4 src pixels
+
+    ; calculate sse
+    uxtb16  r6, r8              ; byte (two pixels) to halfwords
+    uxtb16  r7, r8, ror #8      ; another two pixels to halfwords
+    smlad   r4, r6, r6, r4      ; dual signed multiply, add and accumulate (1)
+
+    ; 4th 4 pixels
+    ldr     r6, [r2, #0xc]      ; load 4 ref pixels
+    smlad   r4, r7, r7, r4      ; dual signed multiply, add and accumulate (2)
+
+    usub8   r8, r5, r6          ; calculate difference
+    add     r0, r0, r1          ; set src_ptr to next row
+    sel     r7, r8, lr          ; select bytes with positive difference
+    usub8   r9, r6, r5          ; calculate difference with reversed operands
+    add     r2, r2, r3          ; set dst_ptr to next row
+    sel     r8, r9, lr          ; select bytes with negative difference
+
+    ; calculate partial sums
+    usad8   r5, r7, lr          ; calculate sum of positive differences
+    usad8   r6, r8, lr          ; calculate sum of negative differences
+    orr     r8, r8, r7          ; differences of all 4 pixels
+
+    subs    r12, r12, #1        ; next row
+
+    ; calculate sse
+    uxtb16  r6, r8              ; byte (two pixels) to halfwords
+    uxtb16  r7, r8, ror #8      ; another two pixels to halfwords
+    smlad   r4, r6, r6, r4      ; dual signed multiply, add and accumulate (1)
+    smlad   r4, r7, r7, r4      ; dual signed multiply, add and accumulate (2)
+
+    bne     loop
+
+    ; return stuff
+    ldr     r1, [sp, #28]       ; get address of sse
+    mov     r0, r4              ; return sse
+    str     r4, [r1]            ; store sse
+
+    pop     {r4-r9, pc}
+
+    ENDP
+
+    END
--- a/vp8/encoder/arm/armv6/vp8_sad16x16_armv6.asm
+++ b/vp8/encoder/arm/armv6/vp8_sad16x16_armv6.asm
@@ -0,0 +1,84 @@
+;
+;  Copyright (c) 2011 The WebM project authors. All Rights Reserved.
+;
+;  Use of this source code is governed by a BSD-style license
+;  that can be found in the LICENSE file in the root of the source
+;  tree. An additional intellectual property rights grant can be found
+;  in the file PATENTS.  All contributing project authors may
+;  be found in the AUTHORS file in the root of the source tree.
+;
+
+
+    EXPORT  |vp8_sad16x16_armv6|
+
+    ARM
+    REQUIRE8
+    PRESERVE8
+
+    AREA ||.text||, CODE, READONLY, ALIGN=2
+
+; r0    const unsigned char *src_ptr
+; r1    int  src_stride
+; r2    const unsigned char *ref_ptr
+; r3    int  ref_stride
+; stack max_sad (not used)
+|vp8_sad16x16_armv6| PROC
+    stmfd   sp!, {r4-r12, lr}
+    mov     r4, #0              ; sad = 0;
+    mov     r5, #8              ; loop count
+
+loop
+    ; 1st row
+    ldr     r6, [r0, #0x0]      ; load 4 src pixels (1A)
+    ldr     r8, [r2, #0x0]      ; load 4 ref pixels (1A)
+    ldr     r7, [r0, #0x4]      ; load 4 src pixels (1A)
+    ldr     r9, [r2, #0x4]      ; load 4 ref pixels (1A)
+    ldr     r10, [r0, #0x8]     ; load 4 src pixels (1B)
+    ldr     r11, [r0, #0xC]     ; load 4 src pixels (1B)
+
+    usada8  r4, r8, r6, r4      ; calculate sad for 4 pixels
+    usad8   r8, r7, r9          ; calculate sad for 4 pixels
+
+    ldr     r12, [r2, #0x8]     ; load 4 ref pixels (1B)
+    ldr     lr, [r2, #0xC]      ; load 4 ref pixels (1B)
+
+    add     r0, r0, r1          ; set src pointer to next row
+    add     r2, r2, r3          ; set dst pointer to next row
+
+    usada8  r4, r10, r12, r4    ; calculate sad for 4 pixels
+    usada8  r8, r11, lr, r8     ; calculate sad for 4 pixels
+
+    ldr     r6, [r0, #0x0]      ; load 4 src pixels (2A)
+    ldr     r7, [r0, #0x4]      ; load 4 src pixels (2A)
+    add     r4, r4, r8          ; add partial sad values
+
+    ; 2nd row
+    ldr     r8, [r2, #0x0]      ; load 4 ref pixels (2A)
+    ldr     r9, [r2, #0x4]      ; load 4 ref pixels (2A)
+    ldr     r10, [r0, #0x8]     ; load 4 src pixels (2B)
+    ldr     r11, [r0, #0xC]     ; load 4 src pixels (2B)
+
+    usada8  r4, r6, r8, r4      ; calculate sad for 4 pixels
+    usad8   r8, r7, r9          ; calculate sad for 4 pixels
+
+    ldr     r12, [r2, #0x8]     ; load 4 ref pixels (2B)
+    ldr     lr, [r2, #0xC]      ; load 4 ref pixels (2B)
+
+    add     r0, r0, r1          ; set src pointer to next row
+    add     r2, r2, r3          ; set dst pointer to next row
+
+    usada8  r4, r10, r12, r4    ; calculate sad for 4 pixels
+    usada8  r8, r11, lr, r8     ; calculate sad for 4 pixels
+
+    subs    r5, r5, #1          ; decrement loop counter
+    add     r4, r4, r8          ; add partial sad values
+
+    bne     loop
+
+    mov     r0, r4              ; return sad
+    ldmfd   sp!, {r4-r12, pc}
+
+    ENDP
+
+    END
+
--- a/vp8/encoder/arm/armv6/vp8_variance8x8_armv6.asm
+++ b/vp8/encoder/arm/armv6/vp8_variance8x8_armv6.asm
@@ -0,0 +1,95 @@
+;
+;  Copyright (c) 2011 The WebM project authors. All Rights Reserved.
+;
+;  Use of this source code is governed by a BSD-style license
+;  that can be found in the LICENSE file in the root of the source
+;  tree. An additional intellectual property rights grant can be found
+;  in the file PATENTS.  All contributing project authors may
+;  be found in the AUTHORS file in the root of the source tree.
+;
+
+
+    EXPORT  |vp8_variance8x8_armv6|
+
+    ARM
+
+    AREA ||.text||, CODE, READONLY, ALIGN=2
+
+; r0    unsigned char *src_ptr
+; r1    int source_stride
+; r2    unsigned char *ref_ptr
+; r3    int  recon_stride
+; stack unsigned int *sse
+|vp8_variance8x8_armv6| PROC
+
+    push    {r4-r10, lr}
+    mov     r12, #8             ; set loop counter to 8 (=block height)
+    mov     r4, #0              ; initialize sum = 0
+    mov     r5, #0              ; initialize sse = 0
+
+loop
+    ; 1st 4 pixels
+    ldr     r6, [r0, #0x0]      ; load 4 src pixels
+    ldr     r7, [r2, #0x0]      ; load 4 ref pixels
+
+    mov     lr, #0              ; constant zero
+
+    usub8   r8, r6, r7          ; calculate difference
+    sel     r10, r8, lr         ; select bytes with positive difference
+    usub8   r9, r7, r6          ; calculate difference with reversed operands
+    sel     r8, r9, lr          ; select bytes with negative difference
+
+    ; calculate partial sums
+    usad8   r6, r10, lr         ; calculate sum of positive differences
+    usad8   r7, r8, lr          ; calculate sum of negative differences
+    orr     r8, r8, r10         ; differences of all 4 pixels
+    ; calculate total sum
+    add    r4, r4, r6           ; add positive differences to sum
+    sub    r4, r4, r7           ; substract negative differences from sum
+
+    ; calculate sse
+    uxtb16  r7, r8              ; byte (two pixels) to halfwords
+    uxtb16  r10, r8, ror #8     ; another two pixels to halfwords
+    smlad   r5, r7, r7, r5      ; dual signed multiply, add and accumulate (1)
+
+    ; 2nd 4 pixels
+    ldr     r6, [r0, #0x4]      ; load 4 src pixels
+    ldr     r7, [r2, #0x4]      ; load 4 ref pixels
+    smlad   r5, r10, r10, r5    ; dual signed multiply, add and accumulate (2)
+
+    usub8   r8, r6, r7          ; calculate difference
+    add     r0, r0, r1          ; set src_ptr to next row
+    sel     r10, r8, lr         ; select bytes with positive difference
+    usub8   r9, r7, r6          ; calculate difference with reversed operands
+    add     r2, r2, r3          ; set dst_ptr to next row
+    sel     r8, r9, lr          ; select bytes with negative difference
+
+    ; calculate partial sums
+    usad8   r6, r10, lr         ; calculate sum of positive differences
+    usad8   r7, r8, lr          ; calculate sum of negative differences
+    orr     r8, r8, r10         ; differences of all 4 pixels
+
+    ; calculate total sum
+    add     r4, r4, r6          ; add positive differences to sum
+    sub     r4, r4, r7          ; substract negative differences from sum
+
+    ; calculate sse
+    uxtb16  r7, r8              ; byte (two pixels) to halfwords
+    uxtb16  r10, r8, ror #8     ; another two pixels to halfwords
+    smlad   r5, r7, r7, r5      ; dual signed multiply, add and accumulate (1)
+    subs    r12, r12, #1        ; next row
+    smlad   r5, r10, r10, r5    ; dual signed multiply, add and accumulate (2)
+
+    bne     loop
+
+    ; return stuff
+    ldr     r8, [sp, #32]       ; get address of sse
+    mul     r1, r4, r4          ; sum * sum
+    str     r5, [r8]            ; store sse
+    sub     r0, r5, r1, ASR #6  ; return (sse - ((sum * sum) >> 6))
+
+    pop     {r4-r10, pc}
+
+    ENDP
+
+    END
--- a/vp8/encoder/arm/boolhuff_arm.c
+++ b/vp8/encoder/arm/boolhuff_arm.c
@@ -9,8 +9,8 @@
 */


-#include "boolhuff.h"
-#include "blockd.h"
+#include "vp8/encoder/boolhuff.h"
+#include "vp8/common/blockd.h"

 const unsigned int vp8_prob_cost[256] =
 {
--- a/vp8/encoder/arm/dct_arm.c
+++ b/vp8/encoder/arm/dct_arm.c
@@ -1,5 +1,5 @@
 /*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *  Copyright (c) 2011 The WebM project authors. All Rights Reserved.
 *
 *  Use of this source code is governed by a BSD-style license
 *  that can be found in the LICENSE file in the root of the source
@@ -8,24 +8,17 @@
 *  be found in the AUTHORS file in the root of the source tree.
 */

+#include "vpx_config.h"
+#include "vp8/encoder/dct.h"

-#if !defined(_mac_specs_h)
-#define _mac_specs_h
+#if HAVE_ARMV6

-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-    extern unsigned int vp8_read_tsc();
-
-    extern unsigned int vp8_get_processor_freq();
-
-    extern unsigned int vpx_has_altivec();
-
-#if defined(__cplusplus)
+void vp8_fast_fdct8x4_armv6(short *input, short *output, int pitch)
+{
+    vp8_fast_fdct4x4_armv6(input,   output,    pitch);
+    vp8_fast_fdct4x4_armv6(input + 4, output + 16, pitch);
 }
-#endif
+
+#endif /* HAVE_ARMV6 */


-#endif
--- a/vp8/encoder/arm/dct_arm.h
+++ b/vp8/encoder/arm/dct_arm.h
@@ -14,12 +14,21 @@

 #if HAVE_ARMV6
 extern prototype_fdct(vp8_short_walsh4x4_armv6);
+extern prototype_fdct(vp8_fast_fdct4x4_armv6);
+extern prototype_fdct(vp8_fast_fdct8x4_armv6);

 #if !CONFIG_RUNTIME_CPU_DETECT
 #undef  vp8_fdct_walsh_short4x4
 #define vp8_fdct_walsh_short4x4 vp8_short_walsh4x4_armv6
+
+#undef  vp8_fdct_fast4x4
+#define vp8_fdct_fast4x4 vp8_fast_fdct4x4_armv6
+
+#undef  vp8_fdct_fast8x4
+#define vp8_fdct_fast8x4 vp8_fast_fdct8x4_armv6
 #endif
-#endif
+
+#endif /* HAVE_ARMV6 */

 #if HAVE_ARMV7
 extern prototype_fdct(vp8_short_fdct4x4_neon);
--- a/vp8/encoder/arm/encodemb_arm.c
+++ b/vp8/encoder/arm/encodemb_arm.c
@@ -9,13 +9,13 @@
 */


-#include "encodemb.h"
-#include "reconinter.h"
-#include "quantize.h"
-#include "invtrans.h"
-#include "recon.h"
-#include "reconintra.h"
-#include "dct.h"
+#include "vp8/encoder/encodemb.h"
+#include "vp8/common/reconinter.h"
+#include "vp8/encoder/quantize.h"
+#include "vp8/common/invtrans.h"
+#include "vp8/common/recon.h"
+#include "vp8/common/reconintra.h"
+#include "vp8/encoder/dct.h"
 #include "vpx_mem/vpx_mem.h"

 extern void vp8_subtract_b_neon_func(short *diff, unsigned char *src, unsigned char *pred, int stride, int pitch);
--- a/vp8/encoder/arm/picklpf_arm.c
+++ b/vp8/encoder/arm/picklpf_arm.c
@@ -9,13 +9,13 @@
 */


-#include "onyxc_int.h"
-#include "onyx_int.h"
-#include "quantize.h"
+#include "vp8/common/onyxc_int.h"
+#include "vp8/encoder/onyx_int.h"
+#include "vp8/encoder/quantize.h"
 #include "vpx_mem/vpx_mem.h"
 #include "vpx_scale/yv12extend.h"
 #include "vpx_scale/vpxscale.h"
-#include "alloccommon.h"
+#include "vp8/common/alloccommon.h"

 extern void vp8_memcpy_neon(unsigned char *dst_ptr, unsigned char *src_ptr, int sz);

--- a/vp8/encoder/arm/quantize_arm.c
+++ b/vp8/encoder/arm/quantize_arm.c
@@ -12,9 +12,8 @@
 #include <math.h>
 #include "vpx_mem/vpx_mem.h"

-#include "quantize.h"
-#include "entropy.h"
-#include "predictdc.h"
+#include "vp8/encoder/quantize.h"
+#include "vp8/common/entropy.h"

 DECLARE_ALIGNED(16, const short, vp8_rvsplus1_default_zig_zag1d[16]) =
 {
--- a/vp8/encoder/arm/quantize_arm.h
+++ b/vp8/encoder/arm/quantize_arm.h
@@ -12,6 +12,16 @@
 #ifndef QUANTIZE_ARM_H
 #define QUANTIZE_ARM_H

+#if HAVE_ARMV6
+
+extern prototype_quantize_block(vp8_fast_quantize_b_armv6);
+
+#undef  vp8_quantize_fastquantb
+#define vp8_quantize_fastquantb vp8_fast_quantize_b_armv6
+
+#endif /* HAVE_ARMV6 */
+
+
 #if HAVE_ARMV7
 extern prototype_quantize_block(vp8_fast_quantize_b_neon);

--- a/vp8/encoder/arm/variance_arm.c
+++ b/vp8/encoder/arm/variance_arm.c
@@ -9,12 +9,40 @@
 */

 #include "vpx_config.h"
-#include "variance.h"
-#include "filter.h"
-#include "arm/bilinearfilter_arm.h"
+#include "vp8/encoder/variance.h"
+#include "vp8/common/filter.h"
+#include "vp8/common/arm/bilinearfilter_arm.h"

 #if HAVE_ARMV6

+unsigned int vp8_sub_pixel_variance8x8_armv6
+(
+    const unsigned char  *src_ptr,
+    int  src_pixels_per_line,
+    int  xoffset,
+    int  yoffset,
+    const unsigned char *dst_ptr,
+    int dst_pixels_per_line,
+    unsigned int *sse
+)
+{
+    unsigned short first_pass[10*8];
+    unsigned char  second_pass[8*8];
+    const short *HFilter, *VFilter;
+
+    HFilter = vp8_bilinear_filters[xoffset];
+    VFilter = vp8_bilinear_filters[yoffset];
+
+    vp8_filter_block2d_bil_first_pass_armv6(src_ptr, first_pass,
+                                            src_pixels_per_line,
+                                            9, 8, HFilter);
+    vp8_filter_block2d_bil_second_pass_armv6(first_pass, second_pass,
+                                             8, 8, 8, VFilter);
+
+    return vp8_variance8x8_armv6(second_pass, 8, dst_ptr,
+                                   dst_pixels_per_line, sse);
+}
+
 unsigned int vp8_sub_pixel_variance16x16_armv6
 (
    const unsigned char  *src_ptr,
@@ -43,7 +71,41 @@ unsigned int vp8_sub_pixel_variance16x16_armv6
                                   dst_pixels_per_line, sse);
 }

-#endif
+unsigned int vp8_variance_halfpixvar16x16_h_armv6(
+    const unsigned char *src_ptr,
+    int  source_stride,
+    const unsigned char *ref_ptr,
+    int  recon_stride,
+    unsigned int *sse)
+{
+    return vp8_sub_pixel_variance16x16_armv6(src_ptr, source_stride, 4, 0,
+                                         ref_ptr, recon_stride, sse);
+}
+
+unsigned int vp8_variance_halfpixvar16x16_v_armv6(
+    const unsigned char *src_ptr,
+    int  source_stride,
+    const unsigned char *ref_ptr,
+    int  recon_stride,
+    unsigned int *sse)
+{
+    return vp8_sub_pixel_variance16x16_armv6(src_ptr, source_stride, 0, 4,
+                                         ref_ptr, recon_stride, sse);
+}
+
+unsigned int vp8_variance_halfpixvar16x16_hv_armv6(
+    const unsigned char *src_ptr,
+    int  source_stride,
+    const unsigned char *ref_ptr,
+    int  recon_stride,
+    unsigned int *sse)
+{
+    return vp8_sub_pixel_variance16x16_armv6(src_ptr, source_stride, 4, 4,
+                                         ref_ptr, recon_stride, sse);
+}
+
+#endif /* HAVE_ARMV6 */
+

 #if HAVE_ARMV7

--- a/vp8/encoder/arm/variance_arm.h
+++ b/vp8/encoder/arm/variance_arm.h
@@ -14,21 +14,50 @@

 #if HAVE_ARMV6

+extern prototype_sad(vp8_sad16x16_armv6);
 extern prototype_variance(vp8_variance16x16_armv6);
+extern prototype_variance(vp8_variance8x8_armv6);
 extern prototype_subpixvariance(vp8_sub_pixel_variance16x16_armv6);
+extern prototype_subpixvariance(vp8_sub_pixel_variance8x8_armv6);
+extern prototype_variance(vp8_variance_halfpixvar16x16_h_armv6);
+extern prototype_variance(vp8_variance_halfpixvar16x16_v_armv6);
+extern prototype_variance(vp8_variance_halfpixvar16x16_hv_armv6);
+extern prototype_variance(vp8_mse16x16_armv6);

 #if !CONFIG_RUNTIME_CPU_DETECT

+#undef  vp8_variance_sad16x16
+#define vp8_variance_sad16x16 vp8_sad16x16_armv6
+
 #undef  vp8_variance_subpixvar16x16
 #define vp8_variance_subpixvar16x16 vp8_sub_pixel_variance16x16_armv6

+#undef  vp8_variance_subpixvar8x8
+#define vp8_variance_subpixvar8x8 vp8_sub_pixel_variance8x8_armv6
+
 #undef  vp8_variance_var16x16
 #define vp8_variance_var16x16 vp8_variance16x16_armv6

+#undef  vp8_variance_mse16x16
+#define vp8_variance_mse16x16 vp8_mse16x16_armv6
+
+#undef  vp8_variance_var8x8
+#define vp8_variance_var8x8 vp8_variance8x8_armv6
+
+#undef  vp8_variance_halfpixvar16x16_h
+#define vp8_variance_halfpixvar16x16_h vp8_variance_halfpixvar16x16_h_armv6
+
+#undef  vp8_variance_halfpixvar16x16_v
+#define vp8_variance_halfpixvar16x16_v vp8_variance_halfpixvar16x16_v_armv6
+
+#undef  vp8_variance_halfpixvar16x16_hv
+#define vp8_variance_halfpixvar16x16_hv vp8_variance_halfpixvar16x16_hv_armv6
+
 #endif /* !CONFIG_RUNTIME_CPU_DETECT */

 #endif /* HAVE_ARMV6 */

+
 #if HAVE_ARMV7
 extern prototype_sad(vp8_sad4x4_neon);
 extern prototype_sad(vp8_sad8x8_neon);
--- a/vp8/encoder/asm_enc_offsets.c
+++ b/vp8/encoder/asm_enc_offsets.c
@@ -65,6 +65,17 @@ DEFINE(TOKENLIST_SZ,                            sizeof(TOKENLIST));

 DEFINE(vp8_common_mb_rows,                      offsetof(VP8_COMMON, mb_rows));

+// offsets from BLOCK structure
+DEFINE(vp8_block_coeff,                         offsetof(BLOCK, coeff));
+DEFINE(vp8_block_quant_fast,                    offsetof(BLOCK, quant_fast));
+DEFINE(vp8_block_round,                         offsetof(BLOCK, round));
+
+// offsets from BLOCKD structure
+DEFINE(vp8_blockd_qcoeff,                       offsetof(BLOCKD, qcoeff));
+DEFINE(vp8_blockd_dqcoeff,                      offsetof(BLOCKD, dqcoeff));
+DEFINE(vp8_blockd_dequant,                      offsetof(BLOCKD, dequant));
+DEFINE(vp8_blockd_eob,                          offsetof(BLOCKD, eob));
+
 // These two sizes are used in vp8cx_pack_tokens.  They are hard coded
 // so if the size changes this will have to be adjusted.
 #if HAVE_ARMV5TE
--- a/vp8/encoder/bitstream.c
+++ b/vp8/encoder/bitstream.c
@@ -9,15 +9,15 @@
 */


-#include "header.h"
+#include "vp8/common/header.h"
 #include "encodemv.h"
-#include "entropymode.h"
-#include "findnearmv.h"
+#include "vp8/common/entropymode.h"
+#include "vp8/common/findnearmv.h"
 #include "mcomp.h"
-#include "systemdependent.h"
+#include "vp8/common/systemdependent.h"
 #include <assert.h>
 #include <stdio.h>
-#include "pragmas.h"
+#include "vp8/common/pragmas.h"
 #include "vpx_mem/vpx_mem.h"
 #include "bitstream.h"

@@ -58,16 +58,6 @@ extern unsigned int active_section;
 int count_mb_seg[4] = { 0, 0, 0, 0 };
 #endif

-#if CONFIG_BIG_ENDIAN
-# define make_endian_16(a)  \
-    (((unsigned int)(a & 0xff)) << 8) | (((unsigned int)(a & 0xff00)) >> 8)
-# define make_endian_32(a)                              \
-    (((unsigned int)(a & 0xff)) << 24)    | (((unsigned int)(a & 0xff00)) << 8) |   \
-    (((unsigned int)(a & 0xff0000)) >> 8) | (((unsigned int)(a & 0xff000000)) >> 24)
-#else
-# define make_endian_16(a)  a
-# define make_endian_32(a)  a
-#endif

 static void update_mode(
    vp8_writer *const w,
@@ -1376,6 +1366,7 @@ void vp8_pack_bitstream(VP8_COMP *cpi, unsigned char *dest, unsigned long *size)
    oh.show_frame = (int) pc->show_frame;
    oh.type = (int)pc->frame_type;
    oh.version = pc->version;
+    oh.first_partition_length_in_bytes = 0;

    mb_feature_data_bits = vp8_mb_feature_data_bits;
    cx_data += 3;
@@ -1392,13 +1383,20 @@ void vp8_pack_bitstream(VP8_COMP *cpi, unsigned char *dest, unsigned long *size)
    // every keyframe send startcode, width, height, scale factor, clamp and color type
    if (oh.type == KEY_FRAME)
    {
+        int v;
+
        // Start / synch code
        cx_data[0] = 0x9D;
        cx_data[1] = 0x01;
        cx_data[2] = 0x2a;

-        *((unsigned short *)(cx_data + 3)) = make_endian_16((pc->horiz_scale << 14) | pc->Width);
-        *((unsigned short *)(cx_data + 5)) = make_endian_16((pc->vert_scale << 14) | pc->Height);
+        v = (pc->horiz_scale << 14) | pc->Width;
+        cx_data[3] = v;
+        cx_data[4] = v >> 8;
+
+        v = (pc->vert_scale << 14) | pc->Height;
+        cx_data[5] = v;
+        cx_data[6] = v >> 8;

        extra_bytes_packed = 7;
        cx_data += extra_bytes_packed ;
@@ -1637,6 +1635,21 @@ void vp8_pack_bitstream(VP8_COMP *cpi, unsigned char *dest, unsigned long *size)

    vp8_stop_encode(bc);

+    oh.first_partition_length_in_bytes = cpi->bc.pos;
+
+    /* update frame tag */
+    {
+        int v = (oh.first_partition_length_in_bytes << 5) |
+                (oh.show_frame << 4) |
+                (oh.version << 1) |
+                oh.type;
+
+        dest[0] = v;
+        dest[1] = v >> 8;
+        dest[2] = v >> 16;
+    }
+
+    *size = VP8_HEADER_SIZE + extra_bytes_packed + cpi->bc.pos;

    if (pc->multi_token_partition != ONE_PARTITION)
    {
@@ -1646,9 +1659,7 @@ void vp8_pack_bitstream(VP8_COMP *cpi, unsigned char *dest, unsigned long *size)

        pack_tokens_into_partitions(cpi, cx_data + bc->pos, num_part, &asize);

-        oh.first_partition_length_in_bytes = cpi->bc.pos;
-
-        *size = cpi->bc.pos + VP8_HEADER_SIZE + asize + extra_bytes_packed;
+        *size += asize;
    }
    else
    {
@@ -1662,23 +1673,9 @@ void vp8_pack_bitstream(VP8_COMP *cpi, unsigned char *dest, unsigned long *size)
            pack_tokens(&cpi->bc2, cpi->tok, cpi->tok_count);

        vp8_stop_encode(&cpi->bc2);
-        oh.first_partition_length_in_bytes = cpi->bc.pos ;
-        *size = cpi->bc2.pos + cpi->bc.pos + VP8_HEADER_SIZE + extra_bytes_packed;
-    }

-#if CONFIG_BIG_ENDIAN
-    {
-        int v = (oh.first_partition_length_in_bytes << 5) |
-                (oh.show_frame << 4) |
-                (oh.version << 1) |
-                oh.type;
-
-        v = make_endian_32(v);
-        vpx_memcpy(dest, &v, 3);
+        *size += cpi->bc2.pos;
    }
-#else
-    vpx_memcpy(dest, &oh, 3);
-#endif
 }

 #ifdef ENTROPY_STATS
--- a/Show More
+++ b/Show More