Release v1.6.0 Khaki Campbell Duck

Change-Id: I08da365dd889093f9919476a02ee96ae9615f140
Fix encoder crashes for odd size input
2016-07-20 18:15:41 -07:00 · 2016-07-20 15:02:13 -07:00 · 2016-06-25 11:40:26 -07:00 · 2016-06-25 11:37:20 -07:00 · 2016-06-25 00:43:40 +00:00 · 2016-06-25 00:36:05 +00:00
481 changed files with 14012 additions and 86093 deletions
--- a/.mailmap
+++ b/.mailmap
@@ -3,6 +3,7 @@ Aℓex Converse <aconverse@google.com>
 Aℓex Converse <aconverse@google.com> <alex.converse@gmail.com>
 Alexis Ballier <aballier@gentoo.org> <alexis.ballier@gmail.com>
 Alpha Lam <hclam@google.com> <hclam@chromium.org>
+Daniele Castagna <dcastagna@chromium.org> <dcastagna@google.com>
 Deb Mukherjee <debargha@google.com>
 Erik Niemeyer <erik.a.niemeyer@intel.com> <erik.a.niemeyer@gmail.com>
 Guillaume Martres <gmartres@google.com> <smarter3@gmail.com>
@@ -13,12 +14,15 @@ Jim Bankoski <jimbankoski@google.com>
 Johann Koenig <johannkoenig@google.com>
 Johann Koenig <johannkoenig@google.com> <johann.koenig@duck.com>
 Johann Koenig <johannkoenig@google.com> <johann.koenig@gmail.com>
+Johann Koenig <johannkoenig@google.com> <johannkoenig@chromium.org>
 John Koleszar <jkoleszar@google.com>
 Joshua Litt <joshualitt@google.com> <joshualitt@chromium.org>
 Marco Paniconi <marpan@google.com>
 Marco Paniconi <marpan@google.com> <marpan@chromium.org>
 Pascal Massimino <pascal.massimino@gmail.com>
 Paul Wilkins <paulwilkins@google.com>
+Peter de Rivaz <peter.derivaz@gmail.com>
+Peter de Rivaz <peter.derivaz@gmail.com> <peter.derivaz@argondesign.com>
 Ralph Giles <giles@xiph.org> <giles@entropywave.com>
 Ralph Giles <giles@xiph.org> <giles@mozilla.com>
 Ronald S. Bultje <rsbultje@gmail.com> <rbultje@google.com>
@@ -26,7 +30,8 @@ Sami Pietilä <samipietila@google.com>
 Tamar Levy <tamar.levy@intel.com>
 Tamar Levy <tamar.levy@intel.com> <levytamar82@gmail.com>
 Tero Rintaluoma <teror@google.com> <tero.rintaluoma@on2.com>
-Timothy B. Terriberry <tterribe@xiph.org> Tim Terriberry <tterriberry@mozilla.com>
+Timothy B. Terriberry <tterribe@xiph.org> <tterriberry@mozilla.com>
 Tom Finegan <tomfinegan@google.com>
 Tom Finegan <tomfinegan@google.com> <tomfinegan@chromium.org>
 Yaowu Xu <yaowu@google.com> <yaowu@xuyaowu.com>
+Yaowu Xu <yaowu@google.com> <Yaowu Xu>
--- a/8
+++ b/8
@@ -24,6 +24,7 @@ changjun.yang <changjun.yang@intel.com>
 Charles 'Buck' Krasic <ckrasic@google.com>
 chm <chm@rock-chips.com>
 Christian Duvivier <cduvivier@google.com>
+Daniele Castagna <dcastagna@chromium.org>
 Daniel Kang <ddkang@google.com>
 Deb Mukherjee <debargha@google.com>
 Dim Temp <dimtemp0@gmail.com>
@@ -56,10 +57,12 @@ James Zern <jzern@google.com>
 Jan Gerber <j@mailb.org>
 Jan Kratochvil <jan.kratochvil@redhat.com>
 Janne Salonen <jsalonen@google.com>
+Jean-Yves Avenard <jyavenard@mozilla.com>
 Jeff Faust <jfaust@google.com>
 Jeff Muizelaar <jmuizelaar@mozilla.com>
 Jeff Petkau <jpet@chromium.org>
 Jia Jia <jia.jia@linaro.org>
+Jian Zhou <zhoujian@google.com>
 Jim Bankoski <jimbankoski@google.com>
 Jingning Han <jingning@google.com>
 Joey Parrish <joeyparrish@google.com>
@@ -74,6 +77,7 @@ Justin Clift <justin@salasaga.org>
 Justin Lebar <justin.lebar@gmail.com>
 KO Myung-Hun <komh@chollian.net>
 Lawrence Velázquez <larryv@macports.org>
+Linfeng Zhang <linfengz@google.com>
 Lou Quillio <louquillio@google.com>
 Luca Barbato <lu_zero@gentoo.org>
 Makoto Kato <makoto.kt@gmail.com>
@@ -107,9 +111,11 @@ Rob Bradford <rob@linux.intel.com>
 Ronald S. Bultje <rsbultje@gmail.com>
 Rui Ueyama <ruiu@google.com>
 Sami Pietilä <samipietila@google.com>
+Sasi Inguva <isasi@google.com>
 Scott Graham <scottmg@chromium.org>
 Scott LaVarnway <slavarnway@google.com>
 Sean McGovern <gseanmcg@gmail.com>
+Sergey Kolomenkin <kolomenkin@gmail.com>
 Sergey Ulanov <sergeyu@chromium.org>
 Shimon Doodkin <helpmepro1@gmail.com>
 Shunyao Li <shunyaoli@google.com>
@@ -126,8 +132,10 @@ Timothy B. Terriberry <tterribe@xiph.org>
 Tom Finegan <tomfinegan@google.com>
 Vignesh Venkatasubramanian <vigneshv@google.com>
 Yaowu Xu <yaowu@google.com>
+Yi Luo <luoyi@google.com>
 Yongzhe Wang <yongzhe@google.com>
 Yunqing Wang <yunqingwang@google.com>
+Yury Gitman <yuryg@google.com>
 Zoe Liu <zoeliu@google.com>
 Google Inc.
 The Mozilla Foundation
--- a/30
+++ b/30
@@ -1,3 +1,33 @@
+2016-07-20 v1.6.0 "Khaki Campbell Duck"
+  This release improves upon the VP9 encoder and speeds up the encoding and
+  decoding processes.
+
+  - Upgrading:
+    This release is ABI incompatible with 1.5.0 due to a new 'color_range' enum
+    in vpx_image and some minor changes to the VP8_COMP structure.
+
+    The default key frame interval for VP9 has changed from 128 to 9999.
+
+  - Enhancement:
+    A core focus has been performance for low end Intel processors. SSSE3
+    instructions such as 'pshufb' have been avoided and instructions have been
+    reordered to better accommodate the more constrained pipelines.
+
+    As a result, devices based on Celeron processors have seen substantial
+    decoding improvements. From Indian Runner Duck to Javan Whistling Duck,
+    decoding speed improved between 10 and 30%. Between Javan Whistling Duck
+    and Khaki Campbell Duck, it improved another 10 to 15%.
+
+    While Celeron benefited most, Core-i5 also improved 5% and 10% between the
+    respective releases.
+
+    Realtime performance for WebRTC for both speed and quality has received a
+    lot of attention.
+
+  - Bug Fixes:
+    A number of fuzzing issues, found variously by Mozilla, Chromium and others,
+    have been fixed and we strongly recommend updating.
+
 2015-11-09 v1.5.0 "Javan Whistling Duck"
  This release improves upon the VP9 encoder and speeds up the encoding and
  decoding processes.
--- a/3
+++ b/3
@@ -1,4 +1,4 @@
-README - 23 March 2015
+README - 20 July 2016

 Welcome to the WebM VP8/VP9 Codec SDK!

@@ -47,7 +47,6 @@ COMPILING THE APPLICATIONS/LIBRARIES:
  --help output of the configure script. As of this writing, the list of
  available targets is:

-    armv6-darwin-gcc
    armv6-linux-rvct
    armv6-linux-gcc
    armv6-none-rvct
--- a/build/make/Android.mk
+++ b/build/make/Android.mk
@@ -174,9 +174,6 @@ endif
 ifeq ($(CONFIG_VP9), yes)
 $$(rtcd_dep_template_SRCS): vp9_rtcd.h
 endif
-ifeq ($(CONFIG_VP10), yes)
-$$(rtcd_dep_template_SRCS): vp10_rtcd.h
-endif
 $$(rtcd_dep_template_SRCS): vpx_scale_rtcd.h
 $$(rtcd_dep_template_SRCS): vpx_dsp_rtcd.h

--- a/build/make/Makefile
+++ b/build/make/Makefile
@@ -119,29 +119,25 @@ utiltest:
 test-no-data-check::
 exampletest-no-data-check utiltest-no-data-check:

-# Add compiler flags for intrinsic files
+# Force to realign stack always on OS/2
 ifeq ($(TOOLCHAIN), x86-os2-gcc)
-STACKREALIGN=-mstackrealign
-else
-STACKREALIGN=
+CFLAGS += -mstackrealign
 endif

 $(BUILD_PFX)%_mmx.c.d: CFLAGS += -mmmx
 $(BUILD_PFX)%_mmx.c.o: CFLAGS += -mmmx
-$(BUILD_PFX)%_sse2.c.d: CFLAGS += -msse2 $(STACKREALIGN)
-$(BUILD_PFX)%_sse2.c.o: CFLAGS += -msse2 $(STACKREALIGN)
-$(BUILD_PFX)%_sse3.c.d: CFLAGS += -msse3 $(STACKREALIGN)
-$(BUILD_PFX)%_sse3.c.o: CFLAGS += -msse3 $(STACKREALIGN)
-$(BUILD_PFX)%_ssse3.c.d: CFLAGS += -mssse3 $(STACKREALIGN)
-$(BUILD_PFX)%_ssse3.c.o: CFLAGS += -mssse3 $(STACKREALIGN)
-$(BUILD_PFX)%_sse4.c.d: CFLAGS += -msse4.1 $(STACKREALIGN)
-$(BUILD_PFX)%_sse4.c.o: CFLAGS += -msse4.1 $(STACKREALIGN)
-$(BUILD_PFX)%_avx.c.d: CFLAGS += -mavx $(STACKREALIGN)
-$(BUILD_PFX)%_avx.c.o: CFLAGS += -mavx $(STACKREALIGN)
-$(BUILD_PFX)%_avx2.c.d: CFLAGS += -mavx2 $(STACKREALIGN)
-$(BUILD_PFX)%_avx2.c.o: CFLAGS += -mavx2 $(STACKREALIGN)
-$(BUILD_PFX)%vp9_reconintra.c.d: CFLAGS += $(STACKREALIGN)
-$(BUILD_PFX)%vp9_reconintra.c.o: CFLAGS += $(STACKREALIGN)
+$(BUILD_PFX)%_sse2.c.d: CFLAGS += -msse2
+$(BUILD_PFX)%_sse2.c.o: CFLAGS += -msse2
+$(BUILD_PFX)%_sse3.c.d: CFLAGS += -msse3
+$(BUILD_PFX)%_sse3.c.o: CFLAGS += -msse3
+$(BUILD_PFX)%_ssse3.c.d: CFLAGS += -mssse3
+$(BUILD_PFX)%_ssse3.c.o: CFLAGS += -mssse3
+$(BUILD_PFX)%_sse4.c.d: CFLAGS += -msse4.1
+$(BUILD_PFX)%_sse4.c.o: CFLAGS += -msse4.1
+$(BUILD_PFX)%_avx.c.d: CFLAGS += -mavx
+$(BUILD_PFX)%_avx.c.o: CFLAGS += -mavx
+$(BUILD_PFX)%_avx2.c.d: CFLAGS += -mavx2
+$(BUILD_PFX)%_avx2.c.o: CFLAGS += -mavx2

 $(BUILD_PFX)%.c.d: %.c
 	$(if $(quiet),@echo "    [DEP] $@")
--- a/build/make/configure.sh
+++ b/build/make/configure.sh
@@ -185,6 +185,7 @@ add_extralibs() {
 #
 # Boolean Manipulation Functions
 #
+
 enable_feature(){
  set_all yes $*
 }
@@ -201,6 +202,20 @@ disabled(){
  eval test "x\$$1" = "xno"
 }

+enable_codec(){
+  enabled "${1}" || echo "  enabling ${1}"
+  enable_feature "${1}"
+
+  is_in "${1}" vp8 vp9 && enable_feature "${1}_encoder" "${1}_decoder"
+}
+
+disable_codec(){
+  disabled "${1}" || echo "  disabling ${1}"
+  disable_feature "${1}"
+
+  is_in "${1}" vp8 vp9 && disable_feature "${1}_encoder" "${1}_decoder"
+}
+
 # Iterates through positional parameters, checks to confirm the parameter has
 # not been explicitly (force) disabled, and enables the setting controlled by
 # the parameter when the setting is not disabled.
@@ -521,22 +536,20 @@ process_common_cmdline() {
        ;;
      --enable-?*|--disable-?*)
        eval `echo "$opt" | sed 's/--/action=/;s/-/ option=/;s/-/_/g'`
-        if echo "${ARCH_EXT_LIST}" | grep "^ *$option\$" >/dev/null; then
+        if is_in ${option} ${ARCH_EXT_LIST}; then
          [ $action = "disable" ] && RTCD_OPTIONS="${RTCD_OPTIONS}--disable-${option} "
        elif [ $action = "disable" ] && ! disabled $option ; then
-          echo "${CMDLINE_SELECT}" | grep "^ *$option\$" >/dev/null ||
-            die_unknown $opt
+          is_in ${option} ${CMDLINE_SELECT} || die_unknown $opt
          log_echo "  disabling $option"
        elif [ $action = "enable" ] && ! enabled $option ; then
-          echo "${CMDLINE_SELECT}" | grep "^ *$option\$" >/dev/null ||
-            die_unknown $opt
+          is_in ${option} ${CMDLINE_SELECT} || die_unknown $opt
          log_echo "  enabling $option"
        fi
        ${action}_feature $option
        ;;
      --require-?*)
        eval `echo "$opt" | sed 's/--/action=/;s/-/ option=/;s/-/_/g'`
-        if echo "${ARCH_EXT_LIST}" none | grep "^ *$option\$" >/dev/null; then
+        if is_in ${option} ${ARCH_EXT_LIST}; then
            RTCD_OPTIONS="${RTCD_OPTIONS}${opt} "
        else
            die_unknown $opt
@@ -638,16 +651,39 @@ show_darwin_sdk_major_version() {
  xcrun --sdk $1 --show-sdk-version 2>/dev/null | cut -d. -f1
 }

+# Print the Xcode version.
+show_xcode_version() {
+  xcodebuild -version | head -n1 | cut -d' ' -f2
+}
+
+# Fails when Xcode version is less than 6.3.
+check_xcode_minimum_version() {
+  xcode_major=$(show_xcode_version | cut -f1 -d.)
+  xcode_minor=$(show_xcode_version | cut -f2 -d.)
+  xcode_min_major=6
+  xcode_min_minor=3
+  if [ ${xcode_major} -lt ${xcode_min_major} ]; then
+    return 1
+  fi
+  if [ ${xcode_major} -eq ${xcode_min_major} ] \
+    && [ ${xcode_minor} -lt ${xcode_min_minor} ]; then
+    return 1
+  fi
+}
+
 process_common_toolchain() {
  if [ -z "$toolchain" ]; then
    gcctarget="${CHOST:-$(gcc -dumpmachine 2> /dev/null)}"

    # detect tgt_isa
    case "$gcctarget" in
+      aarch64*)
+        tgt_isa=arm64
+        ;;
      armv6*)
        tgt_isa=armv6
        ;;
-      armv7*-hardfloat*)
+      armv7*-hardfloat* | armv7*-gnueabihf | arm-*-gnueabihf)
        tgt_isa=armv7
        float_abi=hard
        ;;
@@ -748,7 +784,14 @@ process_common_toolchain() {
  enabled shared && soft_enable pic

  # Minimum iOS version for all target platforms (darwin and iphonesimulator).
-  IOS_VERSION_MIN="6.0"
+  # Shared library framework builds are only possible on iOS 8 and later.
+  if enabled shared; then
+    IOS_VERSION_OPTIONS="--enable-shared"
+    IOS_VERSION_MIN="8.0"
+  else
+    IOS_VERSION_OPTIONS=""
+    IOS_VERSION_MIN="6.0"
+  fi

  # Handle darwin variants. Newer SDKs allow targeting older
  # platforms, so use the newest one available.
@@ -877,7 +920,6 @@ process_common_toolchain() {

      case ${tgt_cc} in
        gcc)
-          CROSS=${CROSS:-arm-none-linux-gnueabi-}
          link_with_cc=gcc
          setup_gnu_toolchain
          arch_int=${tgt_isa##armv}
@@ -899,6 +941,9 @@ EOF
              check_add_cflags -mfpu=neon #-ftree-vectorize
              check_add_asflags -mfpu=neon
            fi
+          elif [ ${tgt_isa} = "arm64" ] || [ ${tgt_isa} = "armv8" ]; then
+            check_add_cflags -march=armv8-a
+            check_add_asflags -march=armv8-a
          else
            check_add_cflags -march=${tgt_isa}
            check_add_asflags -march=${tgt_isa}
@@ -966,6 +1011,10 @@ EOF
          ;;

        android*)
+          if [ -z "${sdk_path}" ]; then
+            die "Must specify --sdk-path for Android builds."
+          fi
+
          SDK_PATH=${sdk_path}
          COMPILER_LOCATION=`find "${SDK_PATH}" \
                             -name "arm-linux-androideabi-gcc*" -print -quit`
@@ -1016,18 +1065,7 @@ EOF
          NM="$(${XCRUN_FIND} nm)"
          RANLIB="$(${XCRUN_FIND} ranlib)"
          AS_SFX=.s
-
-          # Special handling of ld for armv6 because libclang_rt.ios.a does
-          # not contain armv6 support in Apple's clang package:
-          #   Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn).
-          # TODO(tomfinegan): Remove this. Our minimum iOS version (6.0)
-          # renders support for armv6 unnecessary because the 3GS and up
-          # support neon.
-          if [ "${tgt_isa}" = "armv6" ]; then
-            LD="$(${XCRUN_FIND} ld)"
-          else
-            LD="${CXX:-$(${XCRUN_FIND} ld)}"
-          fi
+          LD="${CXX:-$(${XCRUN_FIND} ld)}"

          # ASFLAGS is written here instead of using check_add_asflags
          # because we need to overwrite all of ASFLAGS and purge the
@@ -1053,6 +1091,19 @@ EOF
            [ -d "${try_dir}" ] && add_ldflags -L"${try_dir}"
          done

+          case ${tgt_isa} in
+            armv7|armv7s|armv8|arm64)
+              if enabled neon && ! check_xcode_minimum_version; then
+                soft_disable neon
+                log_echo "  neon disabled: upgrade Xcode (need v6.3+)."
+                if enabled neon_asm; then
+                  soft_disable neon_asm
+                  log_echo "  neon_asm disabled: upgrade Xcode (need v6.3+)."
+                fi
+              fi
+              ;;
+          esac
+
          asm_conversion_cmd="${source_path}/build/make/ads2gas_apple.pl"

          if [ "$(show_darwin_sdk_major_version iphoneos)" -gt 8 ]; then
@@ -1067,7 +1118,7 @@ EOF
          if enabled rvct; then
            # Check if we have CodeSourcery GCC in PATH. Needed for
            # libraries
-            hash arm-none-linux-gnueabi-gcc 2>&- || \
+            which arm-none-linux-gnueabi-gcc 2>&- || \
              die "Couldn't find CodeSourcery GCC from PATH"

            # Use armcc as a linker to enable translation of
@@ -1108,7 +1159,7 @@ EOF
            check_add_ldflags -mfp64
            ;;
          i6400)
-            check_add_cflags -mips64r6 -mabi=64 -funroll-loops -msched-weight 
+            check_add_cflags -mips64r6 -mabi=64 -funroll-loops -msched-weight
            check_add_cflags  -mload-store-pairs -mhard-float -mfp64
            check_add_asflags -mips64r6 -mabi=64 -mhard-float -mfp64
            check_add_ldflags -mips64r6 -mabi=64 -mfp64
@@ -1135,7 +1186,7 @@ EOF
          CC=${CC:-${CROSS}gcc}
          CXX=${CXX:-${CROSS}g++}
          LD=${LD:-${CROSS}gcc}
-          CROSS=${CROSS:-g}
+          CROSS=${CROSS-g}
          ;;
        os2)
          disable_feature pic
@@ -1188,6 +1239,12 @@ EOF
              soft_disable avx2
              ;;
          esac
+          case $vc_version in
+            7|8|9)
+              echo "${tgt_cc} omits stdint.h, disabling webm-io..."
+              soft_disable webm_io
+              ;;
+          esac
          ;;
      esac

--- a/build/make/gen_msvs_proj.sh
+++ b/build/make/gen_msvs_proj.sh
@@ -193,7 +193,7 @@ for opt in "$@"; do
 done

 # Make one call to fix_path for file_list to improve performance.
-fix_file_list
+fix_file_list file_list

 outfile=${outfile:-/dev/stdout}
 guid=${guid:-`generate_uuid`}
--- a/build/make/gen_msvs_vcxproj.sh
+++ b/build/make/gen_msvs_vcxproj.sh
@@ -211,7 +211,7 @@ for opt in "$@"; do
 done

 # Make one call to fix_path for file_list to improve performance.
-fix_file_list
+fix_file_list file_list

 outfile=${outfile:-/dev/stdout}
 guid=${guid:-`generate_uuid`}
--- a/build/make/ios-Info.plist
+++ b/build/make/ios-Info.plist
@@ -0,0 +1,37 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+	<key>CFBundleDevelopmentRegion</key>
+	<string>en</string>
+	<key>CFBundleExecutable</key>
+	<string>VPX</string>
+	<key>CFBundleIdentifier</key>
+	<string>org.webmproject.VPX</string>
+	<key>CFBundleInfoDictionaryVersion</key>
+	<string>6.0</string>
+	<key>CFBundleName</key>
+	<string>VPX</string>
+	<key>CFBundlePackageType</key>
+	<string>FMWK</string>
+	<key>CFBundleShortVersionString</key>
+	<string>${VERSION}</string>
+	<key>CFBundleSignature</key>
+	<string>????</string>
+	<key>CFBundleSupportedPlatforms</key>
+	<array>
+		<string>iPhoneOS</string>
+	</array>
+	<key>CFBundleVersion</key>
+	<string>${VERSION}</string>
+	<key>MinimumOSVersion</key>
+	<string>${IOS_VERSION_MIN}</string>
+	<key>UIDeviceFamily</key>
+	<array>
+		<integer>1</integer>
+		<integer>2</integer>
+	</array>
+	<key>VPXFullVersion</key>
+	<string>${FULLVERSION}</string>
+</dict>
+</plist>
--- a/build/make/iosbuild.sh
+++ b/build/make/iosbuild.sh
@@ -24,6 +24,7 @@ CONFIGURE_ARGS="--disable-docs
                --disable-unit-tests"
 DIST_DIR="_dist"
 FRAMEWORK_DIR="VPX.framework"
+FRAMEWORK_LIB="VPX.framework/VPX"
 HEADER_DIR="${FRAMEWORK_DIR}/Headers/vpx"
 SCRIPT_DIR=$(dirname "$0")
 LIBVPX_SOURCE_DIR=$(cd ${SCRIPT_DIR}/../..; pwd)
@@ -137,6 +138,44 @@ create_vpx_framework_config_shim() {
  printf "#endif  // ${include_guard}" >> "${config_file}"
 }

+# Verifies that $FRAMEWORK_LIB fat library contains requested builds.
+verify_framework_targets() {
+  local requested_cpus=""
+  local cpu=""
+
+  # Extract CPU from full target name.
+  for target; do
+    cpu="${target%%-*}"
+    if [ "${cpu}" = "x86" ]; then
+      # lipo -info outputs i386 for libvpx x86 targets.
+      cpu="i386"
+    fi
+    requested_cpus="${requested_cpus}${cpu} "
+  done
+
+  # Get target CPUs present in framework library.
+  local targets_built=$(${LIPO} -info ${FRAMEWORK_LIB})
+
+  # $LIPO -info outputs a string like the following:
+  #   Architectures in the fat file: $FRAMEWORK_LIB <architectures>
+  # Capture only the architecture strings.
+  targets_built=${targets_built##*: }
+
+  # Sort CPU strings to make the next step a simple string compare.
+  local actual=$(echo ${targets_built} | tr " " "\n" | sort | tr "\n" " ")
+  local requested=$(echo ${requested_cpus} | tr " " "\n" | sort | tr "\n" " ")
+
+  vlog "Requested ${FRAMEWORK_LIB} CPUs: ${requested}"
+  vlog "Actual ${FRAMEWORK_LIB} CPUs: ${actual}"
+
+  if [ "${requested}" != "${actual}" ]; then
+    elog "Actual ${FRAMEWORK_LIB} targets do not match requested target list."
+    elog "  Requested target CPUs: ${requested}"
+    elog "  Actual target CPUs: ${actual}"
+    return 1
+  fi
+}
+
 # Configures and builds each target specified by $1, and then builds
 # VPX.framework.
 build_framework() {
@@ -157,7 +196,12 @@ build_framework() {
  for target in ${targets}; do
    build_target "${target}"
    target_dist_dir="${BUILD_ROOT}/${target}/${DIST_DIR}"
-    lib_list="${lib_list} ${target_dist_dir}/lib/libvpx.a"
+    if [ "${ENABLE_SHARED}" = "yes" ]; then
+      local suffix="dylib"
+    else
+      local suffix="a"
+    fi
+    lib_list="${lib_list} ${target_dist_dir}/lib/libvpx.${suffix}"
  done

  cd "${ORIG_PWD}"
@@ -176,13 +220,25 @@ build_framework() {
  # Copy in vpx_version.h.
  cp -p "${BUILD_ROOT}/${target}/vpx_version.h" "${HEADER_DIR}"

-  vlog "Created fat library ${FRAMEWORK_DIR}/VPX containing:"
+  if [ "${ENABLE_SHARED}" = "yes" ]; then
+    # Adjust the dylib's name so dynamic linking in apps works as expected.
+    install_name_tool -id '@rpath/VPX.framework/VPX' ${FRAMEWORK_DIR}/VPX
+
+    # Copy in Info.plist.
+    cat "${SCRIPT_DIR}/ios-Info.plist" \
+      | sed "s/\${FULLVERSION}/${FULLVERSION}/g" \
+      | sed "s/\${VERSION}/${VERSION}/g" \
+      | sed "s/\${IOS_VERSION_MIN}/${IOS_VERSION_MIN}/g" \
+      > "${FRAMEWORK_DIR}/Info.plist"
+  fi
+
+  # Confirm VPX.framework/VPX contains the targets requested.
+  verify_framework_targets ${targets}
+
+  vlog "Created fat library ${FRAMEWORK_LIB} containing:"
  for lib in ${lib_list}; do
    vlog "  $(echo ${lib} | awk -F / '{print $2, $NF}')"
  done
-
-  # TODO(tomfinegan): Verify that expected targets are included within
-  # VPX.framework/VPX via lipo -info.
 }

 # Trap function. Cleans up the subtree used to build all targets contained in
@@ -213,6 +269,7 @@ iosbuild_usage() {
 cat << EOF
  Usage: ${0##*/} [arguments]
    --help: Display this message and exit.
+    --enable-shared: Build a dynamic framework for use on iOS 8 or later.
    --extra-configure-args <args>: Extra args to pass when configuring libvpx.
    --macosx: Uses darwin15 targets instead of iphonesimulator targets for x86
              and x86_64. Allows linking to framework when builds target MacOSX
@@ -251,6 +308,9 @@ while [ -n "$1" ]; do
      iosbuild_usage
      exit
      ;;
+    --enable-shared)
+      ENABLE_SHARED=yes
+      ;;
    --preserve-build-output)
      PRESERVE_BUILD_OUTPUT=yes
      ;;
@@ -278,6 +338,21 @@ while [ -n "$1" ]; do
  shift
 done

+if [ "${ENABLE_SHARED}" = "yes" ]; then
+  CONFIGURE_ARGS="--enable-shared ${CONFIGURE_ARGS}"
+fi
+
+FULLVERSION=$("${SCRIPT_DIR}"/version.sh --bare "${LIBVPX_SOURCE_DIR}")
+VERSION=$(echo "${FULLVERSION}" | sed -E 's/^v([0-9]+\.[0-9]+\.[0-9]+).*$/\1/')
+
+if [ "$ENABLE_SHARED" = "yes" ]; then
+  IOS_VERSION_OPTIONS="--enable-shared"
+  IOS_VERSION_MIN="8.0"
+else
+  IOS_VERSION_OPTIONS=""
+  IOS_VERSION_MIN="6.0"
+fi
+
 if [ "${VERBOSE}" = "yes" ]; then
 cat << EOF
  BUILD_ROOT=${BUILD_ROOT}
@@ -285,6 +360,7 @@ cat << EOF
  CONFIGURE_ARGS=${CONFIGURE_ARGS}
  EXTRA_CONFIGURE_ARGS=${EXTRA_CONFIGURE_ARGS}
  FRAMEWORK_DIR=${FRAMEWORK_DIR}
+  FRAMEWORK_LIB=${FRAMEWORK_LIB}
  HEADER_DIR=${HEADER_DIR}
  LIBVPX_SOURCE_DIR=${LIBVPX_SOURCE_DIR}
  LIPO=${LIPO}
@@ -292,8 +368,13 @@ cat << EOF
  ORIG_PWD=${ORIG_PWD}
  PRESERVE_BUILD_OUTPUT=${PRESERVE_BUILD_OUTPUT}
  TARGETS="$(print_list "" ${TARGETS})"
+  ENABLE_SHARED=${ENABLE_SHARED}
  OSX_TARGETS="${OSX_TARGETS}"
  SIM_TARGETS="${SIM_TARGETS}"
+  SCRIPT_DIR="${SCRIPT_DIR}"
+  FULLVERSION="${FULLVERSION}"
+  VERSION="${VERSION}"
+  IOS_VERSION_MIN="${IOS_VERSION_MIN}"
 EOF
 fi

--- a/build/make/msvs_common.sh
+++ b/build/make/msvs_common.sh
@@ -39,11 +39,12 @@ fix_path() {
 }

 # Corrects the paths in file_list in one pass for efficiency.
+# $1 is the name of the array to be modified.
 fix_file_list() {
-    # TODO(jzern): this could be more generic and take the array as a param.
-    files=$(fix_path "${file_list[@]}")
+    declare -n array_ref=$1
+    files=$(fix_path "${array_ref[@]}")
    local IFS=$'\n'
-    file_list=($files)
+    array_ref=($files)
 }

 generate_uuid() {
--- a/build/make/version.sh
+++ b/build/make/version.sh
@@ -24,8 +24,9 @@ out_file=${2}
 id=${3:-VERSION_STRING}

 git_version_id=""
-if [ -d "${source_path}/.git" ]; then
+if [ -e "${source_path}/.git" ]; then
    # Source Path is a git working copy. Check for local modifications.
+    # Note that git submodules may have a file as .git, not a directory.
    export GIT_DIR="${source_path}/.git"
    git_version_id=`git describe --match=v[0-9]* 2>/dev/null`
 fi
--- a/64
+++ b/64
@@ -40,7 +40,6 @@ Advanced options:
                                  hardware decoder compatibility
  ${toggle_vp8}                   VP8 codec support
  ${toggle_vp9}                   VP9 codec support
-  ${toggle_vp10}                  VP10 codec support
  ${toggle_internal_stats}        output of encoder internal stats for debug, if supported (encoders)
  ${toggle_postproc}              postprocessing
  ${toggle_vp9_postproc}          vp9 specific postprocessing
@@ -98,11 +97,11 @@ EOF

 # all_platforms is a list of all supported target platforms. Maintain
 # alphabetically by architecture, generic-gnu last.
-all_platforms="${all_platforms} armv6-darwin-gcc"
+all_platforms="${all_platforms} arm64-darwin-gcc"
+all_platforms="${all_platforms} arm64-linux-gcc"
 all_platforms="${all_platforms} armv6-linux-rvct"
 all_platforms="${all_platforms} armv6-linux-gcc"
 all_platforms="${all_platforms} armv6-none-rvct"
-all_platforms="${all_platforms} arm64-darwin-gcc"
 all_platforms="${all_platforms} armv7-android-gcc"   #neon Cortex-A8
 all_platforms="${all_platforms} armv7-darwin-gcc"    #neon Cortex-A8
 all_platforms="${all_platforms} armv7-linux-rvct"    #neon Cortex-A8
@@ -112,6 +111,7 @@ all_platforms="${all_platforms} armv7-win32-vs11"
 all_platforms="${all_platforms} armv7-win32-vs12"
 all_platforms="${all_platforms} armv7-win32-vs14"
 all_platforms="${all_platforms} armv7s-darwin-gcc"
+all_platforms="${all_platforms} armv8-linux-gcc"
 all_platforms="${all_platforms} mips32-linux-gcc"
 all_platforms="${all_platforms} mips64-linux-gcc"
 all_platforms="${all_platforms} sparc-solaris-gcc"
@@ -195,12 +195,8 @@ if [ ${doxy_major:-0} -ge 1 ]; then
 fi

 # disable codecs when their source directory does not exist
-[ -d "${source_path}/vp8" ] || disable_feature vp8
-[ -d "${source_path}/vp9" ] || disable_feature vp9
-[ -d "${source_path}/vp10" ] || disable_feature vp10
-
-# disable vp10 codec by default
-disable_feature vp10
+[ -d "${source_path}/vp8" ] || disable_codec vp8
+[ -d "${source_path}/vp9" ] || disable_codec vp9

 # install everything except the sources, by default. sources will have
 # to be enabled when doing dist builds, since that's no longer a common
@@ -222,13 +218,10 @@ CODECS="
    vp8_decoder
    vp9_encoder
    vp9_decoder
-    vp10_encoder
-    vp10_decoder
 "
 CODEC_FAMILIES="
    vp8
    vp9
-    vp10
 "

 ARCH_LIST="
@@ -263,9 +256,7 @@ ARCH_EXT_LIST="
 HAVE_LIST="
    ${ARCH_EXT_LIST}
    vpx_ports
-    stdint_h
    pthread_h
-    sys_mman_h
    unistd_h
 "
 EXPERIMENT_LIST="
@@ -393,15 +384,19 @@ process_cmdline() {
    for opt do
        optval="${opt#*=}"
        case "$opt" in
-        --disable-codecs) for c in ${CODECS}; do disable_feature $c; done ;;
+        --disable-codecs)
+          for c in ${CODEC_FAMILIES}; do disable_codec $c; done
+          ;;
        --enable-?*|--disable-?*)
        eval `echo "$opt" | sed 's/--/action=/;s/-/ option=/;s/-/_/g'`
-        if echo "${EXPERIMENT_LIST}" | grep "^ *$option\$" >/dev/null; then
+        if is_in ${option} ${EXPERIMENT_LIST}; then
            if enabled experimental; then
                ${action}_feature $option
            else
                log_echo "Ignoring $opt -- not in experimental mode."
            fi
+        elif is_in ${option} "${CODECS} ${CODEC_FAMILIES}"; then
+            ${action}_codec ${option}
        else
            process_common_cmdline $opt
        fi
@@ -415,14 +410,6 @@ process_cmdline() {
 post_process_cmdline() {
    c=""

-    # If the codec family is disabled, disable all components of that family.
-    # If the codec family is enabled, enable all components of that family.
-    log_echo "Configuring selected codecs"
-    for c in ${CODECS}; do
-        disabled ${c%%_*} && disable_feature ${c}
-        enabled ${c%%_*} && enable_feature ${c}
-    done
-
    # Enable all detected codecs, if they haven't been disabled
    for c in ${CODECS}; do soft_enable $c; done

@@ -517,13 +504,18 @@ process_detect() {
        # Can only build shared libs on a subset of platforms. Doing this check
        # here rather than at option parse time because the target auto-detect
        # magic happens after the command line has been parsed.
-        if ! enabled linux && ! enabled os2; then
+        case "${tgt_os}" in
+        linux|os2|darwin*|iphonesimulator*)
+            # Supported platforms
+            ;;
+        *)
            if enabled gnu; then
                echo "--enable-shared is only supported on ELF; assuming this is OK"
            else
-                die "--enable-shared only supported on ELF and OS/2 for now"
+                die "--enable-shared only supported on ELF, OS/2, and Darwin for now"
            fi
-        fi
+            ;;
+        esac
    fi
    if [ -z "$CC" ] || enabled external_build; then
        echo "Bypassing toolchain for environment detection."
@@ -550,16 +542,12 @@ process_detect() {
            # Specialize windows and POSIX environments.
            case $toolchain in
                *-win*-*)
-                    case $header-$toolchain in
-                        stdint*-gcc) true;;
-                        *) false;;
-                    esac && enable_feature $var
-                    ;;
+                    # Don't check for any headers in Windows builds.
+                    false
+                ;;
                *)
                    case $header in
-                        stdint.h) true;;
                        pthread.h) true;;
-                        sys/mman.h) true;;
                        unistd.h) true;;
                        *) false;;
                    esac && enable_feature $var
@@ -575,9 +563,7 @@ process_detect() {
 int main(void) {return 0;}
 EOF
    # check system headers
-    check_header stdint.h
    check_header pthread.h
-    check_header sys/mman.h
    check_header unistd.h # for sysconf(3) and friends.

    check_header vpx/vpx_integer.h -I${source_path} && enable_feature vpx_ports
@@ -608,7 +594,11 @@ process_toolchain() {
          ;;
          *) check_add_cflags -Wunused-but-set-variable ;;
        esac
-        enabled extra_warnings || check_add_cflags -Wno-unused-function
+        if enabled mips || [ -z "${INLINE}" ]; then
+          enabled extra_warnings || check_add_cflags -Wno-unused-function
+        else
+          check_add_cflags -Wunused-function
+        fi
    fi

    if enabled icc; then
--- a/examples.mk
+++ b/examples.mk
@@ -36,21 +36,30 @@ LIBYUV_SRCS +=  third_party/libyuv/include/libyuv/basic_types.h  \
                third_party/libyuv/source/scale_neon64.cc \
                third_party/libyuv/source/scale_win.cc \

-LIBWEBM_COMMON_SRCS += third_party/libwebm/webmids.hpp
+LIBWEBM_COMMON_SRCS += third_party/libwebm/common/hdr_util.cc \
+                       third_party/libwebm/common/hdr_util.h \
+                       third_party/libwebm/common/webmids.h

-LIBWEBM_MUXER_SRCS += third_party/libwebm/mkvmuxer.cpp \
-                      third_party/libwebm/mkvmuxerutil.cpp \
-                      third_party/libwebm/mkvwriter.cpp \
-                      third_party/libwebm/mkvmuxer.hpp \
-                      third_party/libwebm/mkvmuxertypes.hpp \
-                      third_party/libwebm/mkvmuxerutil.hpp \
-                      third_party/libwebm/mkvparser.hpp \
-                      third_party/libwebm/mkvwriter.hpp
+LIBWEBM_MUXER_SRCS += third_party/libwebm/mkvmuxer/mkvmuxer.cc \
+                      third_party/libwebm/mkvmuxer/mkvmuxerutil.cc \
+                      third_party/libwebm/mkvmuxer/mkvwriter.cc \
+                      third_party/libwebm/mkvmuxer/mkvmuxer.h \
+                      third_party/libwebm/mkvmuxer/mkvmuxertypes.h \
+                      third_party/libwebm/mkvmuxer/mkvmuxerutil.h \
+                      third_party/libwebm/mkvparser/mkvparser.h \
+                      third_party/libwebm/mkvmuxer/mkvwriter.h
+
+LIBWEBM_PARSER_SRCS = third_party/libwebm/mkvparser/mkvparser.cc \
+                      third_party/libwebm/mkvparser/mkvreader.cc \
+                      third_party/libwebm/mkvparser/mkvparser.h \
+                      third_party/libwebm/mkvparser/mkvreader.h
+
+# Add compile flags and include path for libwebm sources.
+ifeq ($(CONFIG_WEBM_IO),yes)
+  CXXFLAGS     += -D__STDC_CONSTANT_MACROS -D__STDC_LIMIT_MACROS
+  INC_PATH-yes += $(SRC_PATH_BARE)/third_party/libwebm
+endif

-LIBWEBM_PARSER_SRCS = third_party/libwebm/mkvparser.cpp \
-                      third_party/libwebm/mkvreader.cpp \
-                      third_party/libwebm/mkvparser.hpp \
-                      third_party/libwebm/mkvreader.hpp

 # List of examples to build. UTILS are tools meant for distribution
 # while EXAMPLES demonstrate specific portions of the API.
@@ -70,6 +79,7 @@ ifeq ($(CONFIG_LIBYUV),yes)
 endif
 ifeq ($(CONFIG_WEBM_IO),yes)
  vpxdec.SRCS                 += $(LIBWEBM_COMMON_SRCS)
+  vpxdec.SRCS                 += $(LIBWEBM_MUXER_SRCS)
  vpxdec.SRCS                 += $(LIBWEBM_PARSER_SRCS)
  vpxdec.SRCS                 += webmdec.cc webmdec.h
 endif
@@ -93,6 +103,7 @@ endif
 ifeq ($(CONFIG_WEBM_IO),yes)
  vpxenc.SRCS                 += $(LIBWEBM_COMMON_SRCS)
  vpxenc.SRCS                 += $(LIBWEBM_MUXER_SRCS)
+  vpxenc.SRCS                 += $(LIBWEBM_PARSER_SRCS)
  vpxenc.SRCS                 += webmenc.cc webmenc.h
 endif
 vpxenc.GUID                  = 548DEC74-7A15-4B2B-AFC3-AA102E7C25C1
--- a/examples/simple_encoder.c
+++ b/examples/simple_encoder.c
@@ -109,8 +109,8 @@ static const char *exec_name;
 void usage_exit(void) {
  fprintf(stderr,
          "Usage: %s <codec> <width> <height> <infile> <outfile> "
-              "<keyframe-interval> [<error-resilient>]\nSee comments in "
-              "simple_encoder.c for more information.\n",
+              "<keyframe-interval> <error-resilient> <frames to encode>\n"
+              "See comments in simple_encoder.c for more information.\n",
          exec_name);
  exit(EXIT_FAILURE);
 }
@@ -147,6 +147,7 @@ static int encode_frame(vpx_codec_ctx_t *codec,
  return got_pkts;
 }

+// TODO(tomfinegan): Improve command line parsing and add args for bitrate/fps.
 int main(int argc, char **argv) {
  FILE *infile = NULL;
  vpx_codec_ctx_t codec;
@@ -157,12 +158,11 @@ int main(int argc, char **argv) {
  VpxVideoInfo info = {0};
  VpxVideoWriter *writer = NULL;
  const VpxInterface *encoder = NULL;
-  const int fps = 30;        // TODO(dkovalev) add command line argument
-  const int bitrate = 200;   // kbit/s TODO(dkovalev) add command line argument
+  const int fps = 30;
+  const int bitrate = 200;
  int keyframe_interval = 0;
-
-  // TODO(dkovalev): Add some simple command line parsing code to make the
-  // command line more flexible.
+  int max_frames = 0;
+  int frames_encoded = 0;
  const char *codec_arg = NULL;
  const char *width_arg = NULL;
  const char *height_arg = NULL;
@@ -172,7 +172,7 @@ int main(int argc, char **argv) {

  exec_name = argv[0];

-  if (argc < 7)
+  if (argc != 9)
    die("Invalid number of arguments");

  codec_arg = argv[1];
@@ -181,6 +181,7 @@ int main(int argc, char **argv) {
  infile_arg = argv[4];
  outfile_arg = argv[5];
  keyframe_interval_arg = argv[6];
+  max_frames = strtol(argv[8], NULL, 0);

  encoder = get_vpx_encoder_by_name(codec_arg);
  if (!encoder)
@@ -219,7 +220,7 @@ int main(int argc, char **argv) {
  cfg.g_timebase.num = info.time_base.numerator;
  cfg.g_timebase.den = info.time_base.denominator;
  cfg.rc_target_bitrate = bitrate;
-  cfg.g_error_resilient = argc > 7 ? strtol(argv[7], NULL, 0) : 0;
+  cfg.g_error_resilient = strtol(argv[7], NULL, 0);

  writer = vpx_video_writer_open(outfile_arg, kContainerIVF, &info);
  if (!writer)
@@ -237,6 +238,9 @@ int main(int argc, char **argv) {
    if (keyframe_interval > 0 && frame_count % keyframe_interval == 0)
      flags |= VPX_EFLAG_FORCE_KF;
    encode_frame(&codec, &raw, frame_count++, flags, writer);
+    frames_encoded++;
+    if (max_frames > 0 && frames_encoded >= max_frames)
+      break;
  }

  // Flush encoder.
--- a/examples/twopass_encoder.c
+++ b/examples/twopass_encoder.c
@@ -59,7 +59,9 @@
 static const char *exec_name;

 void usage_exit(void) {
-  fprintf(stderr, "Usage: %s <codec> <width> <height> <infile> <outfile>\n",
+  fprintf(stderr,
+          "Usage: %s <codec> <width> <height> <infile> <outfile> "
+              "<frame limit>\n",
          exec_name);
  exit(EXIT_FAILURE);
 }
@@ -129,7 +131,8 @@ static int encode_frame(vpx_codec_ctx_t *ctx,
 static vpx_fixed_buf_t pass0(vpx_image_t *raw,
                             FILE *infile,
                             const VpxInterface *encoder,
-                             const vpx_codec_enc_cfg_t *cfg) {
+                             const vpx_codec_enc_cfg_t *cfg,
+                             int max_frames) {
  vpx_codec_ctx_t codec;
  int frame_count = 0;
  vpx_fixed_buf_t stats = {NULL, 0};
@@ -142,6 +145,8 @@ static vpx_fixed_buf_t pass0(vpx_image_t *raw,
    ++frame_count;
    get_frame_stats(&codec, raw, frame_count, 1, 0, VPX_DL_GOOD_QUALITY,
                    &stats);
+    if (max_frames > 0 && frame_count >= max_frames)
+      break;
  }

  // Flush encoder.
@@ -159,7 +164,8 @@ static void pass1(vpx_image_t *raw,
                  FILE *infile,
                  const char *outfile_name,
                  const VpxInterface *encoder,
-                  const vpx_codec_enc_cfg_t *cfg) {
+                  const vpx_codec_enc_cfg_t *cfg,
+                  int max_frames) {
  VpxVideoInfo info = {
    encoder->fourcc,
    cfg->g_w,
@@ -181,6 +187,9 @@ static void pass1(vpx_image_t *raw,
  while (vpx_img_read(raw, infile)) {
    ++frame_count;
    encode_frame(&codec, raw, frame_count, 1, 0, VPX_DL_GOOD_QUALITY, writer);
+
+    if (max_frames > 0 && frame_count >= max_frames)
+      break;
  }

  // Flush encoder.
@@ -213,11 +222,14 @@ int main(int argc, char **argv) {
  const char *const height_arg = argv[3];
  const char *const infile_arg = argv[4];
  const char *const outfile_arg = argv[5];
+  int max_frames = 0;
  exec_name = argv[0];

-  if (argc != 6)
+  if (argc != 7)
    die("Invalid number of arguments.");

+  max_frames = strtol(argv[6], NULL, 0);
+
  encoder = get_vpx_encoder_by_name(codec_arg);
  if (!encoder)
    die("Unsupported codec.");
@@ -249,13 +261,13 @@ int main(int argc, char **argv) {

  // Pass 0
  cfg.g_pass = VPX_RC_FIRST_PASS;
-  stats = pass0(&raw, infile, encoder, &cfg);
+  stats = pass0(&raw, infile, encoder, &cfg, max_frames);

  // Pass 1
  rewind(infile);
  cfg.g_pass = VPX_RC_LAST_PASS;
  cfg.rc_twopass_stats_in = stats;
-  pass1(&raw, infile, outfile_arg, encoder, &cfg);
+  pass1(&raw, infile, outfile_arg, encoder, &cfg, max_frames);
  free(stats.buf);

  vpx_img_free(&raw);
--- a/examples/vp8_multi_resolution_encoder.c
+++ b/examples/vp8_multi_resolution_encoder.c
@@ -29,13 +29,6 @@
 #include <math.h>
 #include <assert.h>
 #include <sys/time.h>
-#if USE_POSIX_MMAP
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <sys/mman.h>
-#include <fcntl.h>
-#include <unistd.h>
-#endif
 #include "vpx_ports/vpx_timer.h"
 #include "vpx/vpx_encoder.h"
 #include "vpx/vp8cx.h"
@@ -354,8 +347,7 @@ int main(int argc, char **argv)
    double               psnr_totals[NUM_ENCODERS][4] = {{0,0}};
    int                  psnr_count[NUM_ENCODERS] = {0};

-    double               cx_time = 0;
-    struct  timeval      tv1, tv2, difftv;
+    int64_t              cx_time = 0;

    /* Set the required target bitrates for each resolution level.
     * If target bitrate for highest-resolution level is set to 0,
@@ -589,6 +581,7 @@ int main(int argc, char **argv)

    while(frame_avail || got_data)
    {
+        struct vpx_usec_timer timer;
        vpx_codec_iter_t iter[NUM_ENCODERS]={NULL};
        const vpx_codec_cx_pkt_t *pkt[NUM_ENCODERS];

@@ -643,18 +636,18 @@ int main(int argc, char **argv)
            vpx_codec_control(&codec[i], VP8E_SET_TEMPORAL_LAYER_ID, layer_id);
        }

-        gettimeofday(&tv1, NULL);
        /* Encode each frame at multi-levels */
        /* Note the flags must be set to 0 in the encode call if they are set
           for each frame with the vpx_codec_control(), as done above. */
+        vpx_usec_timer_start(&timer);
        if(vpx_codec_encode(&codec[0], frame_avail? &raw[0] : NULL,
            frame_cnt, 1, 0, arg_deadline))
        {
            die_codec(&codec[0], "Failed to encode frame");
        }
-        gettimeofday(&tv2, NULL);
-        timersub(&tv2, &tv1, &difftv);
-        cx_time += (double)(difftv.tv_sec * 1000000 + difftv.tv_usec);
+        vpx_usec_timer_mark(&timer);
+        cx_time += vpx_usec_timer_elapsed(&timer);
+
        for (i=NUM_ENCODERS-1; i>=0 ; i--)
        {
            got_data = 0;
@@ -693,8 +686,10 @@ int main(int argc, char **argv)
        frame_cnt++;
    }
    printf("\n");
-    printf("FPS for encoding %d %f %f \n", frame_cnt, (float)cx_time / 1000000,
-           1000000 * (double)frame_cnt / (double)cx_time);
+    printf("Frame cnt and encoding time/FPS stats for encoding: %d %f %f \n",
+            frame_cnt,
+            1000 * (float)cx_time / (double)(frame_cnt * 1000000),
+            1000000 * (double)frame_cnt / (double)cx_time);

    fclose(infile);

--- a/examples/vp9_spatial_svc_encoder.c
+++ b/examples/vp9_spatial_svc_encoder.c
@@ -30,6 +30,7 @@
 #include "vpx/vp8cx.h"
 #include "vpx/vpx_encoder.h"
 #include "../vpxstats.h"
+#include "vp9/encoder/vp9_encoder.h"
 #define OUTPUT_RC_STATS 1

 static const arg_def_t skip_frames_arg =
@@ -749,6 +750,7 @@ int main(int argc, const char **argv) {
    cx_time += vpx_usec_timer_elapsed(&timer);

    printf("%s", vpx_svc_get_message(&svc_ctx));
+    fflush(stdout);
    if (res != VPX_CODEC_OK) {
      die_codec(&codec, "Failed to encode frame");
    }
@@ -756,6 +758,7 @@ int main(int argc, const char **argv) {
    while ((cx_pkt = vpx_codec_get_cx_data(&codec, &iter)) != NULL) {
      switch (cx_pkt->kind) {
        case VPX_CODEC_CX_FRAME_PKT: {
+          SvcInternal_t *const si = (SvcInternal_t *)svc_ctx.internal;
          if (cx_pkt->data.frame.sz > 0) {
 #if OUTPUT_RC_STATS
            uint32_t sizes[8];
@@ -851,6 +854,8 @@ int main(int argc, const char **argv) {
          printf("SVC frame: %d, kf: %d, size: %d, pts: %d\n", frames_received,
                 !!(cx_pkt->data.frame.flags & VPX_FRAME_IS_KEY),
                 (int)cx_pkt->data.frame.sz, (int)cx_pkt->data.frame.pts);
+          if (enc_cfg.ss_number_layers == 1 && enc_cfg.ts_number_layers == 1)
+            si->bytes_sum[0] += (int)cx_pkt->data.frame.sz;
          ++frames_received;
          break;
        }
--- a/examples/vpx_temporal_svc_encoder.c
+++ b/examples/vpx_temporal_svc_encoder.c
@@ -41,7 +41,7 @@ enum denoiserState {
  kDenoiserOnAdaptive
 };

-static int mode_to_num_layers[12] = {1, 2, 2, 3, 3, 3, 3, 5, 2, 3, 3, 3};
+static int mode_to_num_layers[13] = {1, 2, 2, 3, 3, 3, 3, 5, 2, 3, 3, 3, 3};

 // For rate control encoding stats.
 struct RateControlMetrics {
@@ -432,7 +432,32 @@ static void set_temporal_layer_pattern(int layering_mode,
      layer_flags[7] = layer_flags[3];
      break;
    }
-    case 11:
+    case 11: {
+      // 3-layers structure with one reference frame.
+      // This works same as temporal_layering_mode 3.
+      // This was added to compare with vp9_spatial_svc_encoder.
+
+      // 3-layers, 4-frame period.
+      int ids[4] = {0, 2, 1, 2};
+      cfg->ts_periodicity = 4;
+      *flag_periodicity = 4;
+      cfg->ts_number_layers = 3;
+      cfg->ts_rate_decimator[0] = 4;
+      cfg->ts_rate_decimator[1] = 2;
+      cfg->ts_rate_decimator[2] = 1;
+      memcpy(cfg->ts_layer_id, ids, sizeof(ids));
+      // 0=L, 1=GF, 2=ARF, Intra-layer prediction disabled.
+      layer_flags[0] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF |
+          VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF;
+      layer_flags[2] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF |
+          VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST;
+      layer_flags[1] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF |
+          VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF;
+      layer_flags[3] = VP8_EFLAG_NO_REF_LAST | VP8_EFLAG_NO_REF_ARF |
+          VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF;
+      break;
+    }
+    case 12:
    default: {
      // 3-layers structure as in case 10, but no sync/refresh points for
      // layer 1 and 2.
@@ -530,7 +555,7 @@ int main(int argc, char **argv) {
  }

  layering_mode = strtol(argv[10], NULL, 0);
-  if (layering_mode < 0 || layering_mode > 12) {
+  if (layering_mode < 0 || layering_mode > 13) {
    die("Invalid layering mode (0..12) %s", argv[10]);
  }

@@ -690,7 +715,7 @@ int main(int argc, char **argv) {
    vpx_codec_control(&codec, VP8E_SET_CPUUSED, speed);
    vpx_codec_control(&codec, VP9E_SET_AQ_MODE, 3);
    vpx_codec_control(&codec, VP9E_SET_FRAME_PERIODIC_BOOST, 0);
-    vpx_codec_control(&codec, VP9E_SET_NOISE_SENSITIVITY, 0);
+    vpx_codec_control(&codec, VP9E_SET_NOISE_SENSITIVITY, kDenoiserOff);
    vpx_codec_control(&codec, VP8E_SET_STATIC_THRESHOLD, 1);
    vpx_codec_control(&codec, VP9E_SET_TUNE_CONTENT, 0);
    vpx_codec_control(&codec, VP9E_SET_TILE_COLUMNS, (cfg.g_threads >> 1));
--- a/ivfdec.c
+++ b/ivfdec.c
@@ -23,7 +23,7 @@ static void fix_framerate(int *num, int *den) {
  // we can guess the framerate using only the timebase in this
  // case. Other files would require reading ahead to guess the
  // timebase, like we do for webm.
-  if (*num < 1000) {
+  if (*den > 0 && *den < 1000000000 && *num > 0 && *num < 1000) {
    // Correct for the factor of 2 applied to the timebase in the encoder.
    if (*num & 1)
      *den *= 2;
--- a/libs.mk
+++ b/libs.mk
@@ -109,40 +109,6 @@ endif
 VP9_PREFIX=vp9/
 $(BUILD_PFX)$(VP9_PREFIX)%.c.o: CFLAGS += -Wextra

-#  VP10 make file
-ifeq ($(CONFIG_VP10),yes)
-  VP10_PREFIX=vp10/
-  include $(SRC_PATH_BARE)/$(VP10_PREFIX)vp10_common.mk
-endif
-
-ifeq ($(CONFIG_VP10_ENCODER),yes)
-  VP10_PREFIX=vp10/
-  include $(SRC_PATH_BARE)/$(VP10_PREFIX)vp10cx.mk
-  CODEC_SRCS-yes += $(addprefix $(VP10_PREFIX),$(call enabled,VP10_CX_SRCS))
-  CODEC_EXPORTS-yes += $(addprefix $(VP10_PREFIX),$(VP10_CX_EXPORTS))
-  CODEC_SRCS-yes += $(VP10_PREFIX)vp10cx.mk vpx/vp8.h vpx/vp8cx.h
-  INSTALL-LIBS-yes += include/vpx/vp8.h include/vpx/vp8cx.h
-  INSTALL-LIBS-$(CONFIG_SPATIAL_SVC) += include/vpx/svc_context.h
-  INSTALL_MAPS += include/vpx/% $(SRC_PATH_BARE)/$(VP10_PREFIX)/%
-  CODEC_DOC_SRCS += vpx/vp8.h vpx/vp8cx.h
-  CODEC_DOC_SECTIONS += vp9 vp9_encoder
-endif
-
-ifeq ($(CONFIG_VP10_DECODER),yes)
-  VP10_PREFIX=vp10/
-  include $(SRC_PATH_BARE)/$(VP10_PREFIX)vp10dx.mk
-  CODEC_SRCS-yes += $(addprefix $(VP10_PREFIX),$(call enabled,VP10_DX_SRCS))
-  CODEC_EXPORTS-yes += $(addprefix $(VP10_PREFIX),$(VP10_DX_EXPORTS))
-  CODEC_SRCS-yes += $(VP10_PREFIX)vp10dx.mk vpx/vp8.h vpx/vp8dx.h
-  INSTALL-LIBS-yes += include/vpx/vp8.h include/vpx/vp8dx.h
-  INSTALL_MAPS += include/vpx/% $(SRC_PATH_BARE)/$(VP10_PREFIX)/%
-  CODEC_DOC_SRCS += vpx/vp8.h vpx/vp8dx.h
-  CODEC_DOC_SECTIONS += vp9 vp9_decoder
-endif
-
-VP10_PREFIX=vp10/
-$(BUILD_PFX)$(VP10_PREFIX)%.c.o: CFLAGS += -Wextra
-
 ifeq ($(CONFIG_ENCODERS),yes)
  CODEC_DOC_SECTIONS += encoder
 endif
@@ -183,6 +149,9 @@ INSTALL-SRCS-$(CONFIG_CODEC_SRCS) += third_party/x86inc/x86inc.asm
 endif
 CODEC_EXPORTS-yes += vpx/exports_com
 CODEC_EXPORTS-$(CONFIG_ENCODERS) += vpx/exports_enc
+ifeq ($(CONFIG_SPATIAL_SVC),yes)
+CODEC_EXPORTS-$(CONFIG_ENCODERS) += vpx/exports_spatial_svc
+endif
 CODEC_EXPORTS-$(CONFIG_DECODERS) += vpx/exports_dec

 INSTALL-LIBS-yes += include/vpx/vpx_codec.h
@@ -260,7 +229,7 @@ OBJS-yes += $(LIBVPX_OBJS)
 LIBS-$(if yes,$(CONFIG_STATIC)) += $(BUILD_PFX)libvpx.a $(BUILD_PFX)libvpx_g.a
 $(BUILD_PFX)libvpx_g.a: $(LIBVPX_OBJS)

-SO_VERSION_MAJOR := 3
+SO_VERSION_MAJOR := 4
 SO_VERSION_MINOR := 0
 SO_VERSION_PATCH := 0
 ifeq ($(filter darwin%,$(TGT_OS)),$(TGT_OS))
@@ -270,6 +239,12 @@ EXPORT_FILE             := libvpx.syms
 LIBVPX_SO_SYMLINKS      := $(addprefix $(LIBSUBDIR)/, \
                             libvpx.dylib  )
 else
+ifeq ($(filter iphonesimulator%,$(TGT_OS)),$(TGT_OS))
+LIBVPX_SO               := libvpx.$(SO_VERSION_MAJOR).dylib
+SHARED_LIB_SUF          := .dylib
+EXPORT_FILE             := libvpx.syms
+LIBVPX_SO_SYMLINKS      := $(addprefix $(LIBSUBDIR)/, libvpx.dylib)
+else
 ifeq ($(filter os2%,$(TGT_OS)),$(TGT_OS))
 LIBVPX_SO               := libvpx$(SO_VERSION_MAJOR).dll
 SHARED_LIB_SUF          := _dll.a
@@ -285,6 +260,7 @@ LIBVPX_SO_SYMLINKS      := $(addprefix $(LIBSUBDIR)/, \
                             libvpx.so.$(SO_VERSION_MAJOR).$(SO_VERSION_MINOR))
 endif
 endif
+endif

 LIBS-$(CONFIG_SHARED) += $(BUILD_PFX)$(LIBVPX_SO)\
                           $(notdir $(LIBVPX_SO_SYMLINKS)) \
@@ -394,6 +370,12 @@ $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)vpx_config.asm
 $(shell $(SRC_PATH_BARE)/build/make/version.sh "$(SRC_PATH_BARE)" $(BUILD_PFX)vpx_version.h)
 CLEAN-OBJS += $(BUILD_PFX)vpx_version.h

+#
+# Add include path for libwebm sources.
+#
+ifeq ($(CONFIG_WEBM_IO),yes)
+  CXXFLAGS += -I$(SRC_PATH_BARE)/third_party/libwebm
+endif

 ##
 ## libvpx test directives
@@ -469,6 +451,7 @@ test_libvpx.$(VCPROJ_SFX): $(LIBVPX_TEST_SRCS) vpx.$(VCPROJ_SFX) gtest.$(VCPROJ_
            $(if $(CONFIG_STATIC_MSVCRT),--static-crt) \
            --out=$@ $(INTERNAL_CFLAGS) $(CFLAGS) \
            -I. -I"$(SRC_PATH_BARE)/third_party/googletest/src/include" \
+            $(if $(CONFIG_WEBM_IO),-I"$(SRC_PATH_BARE)/third_party/libwebm") \
            -L. -l$(CODEC_LIB) -l$(GTEST_LIB) $^

 PROJECTS-$(CONFIG_MSVS) += test_libvpx.$(VCPROJ_SFX)
--- a/md5_utils.c
+++ b/md5_utils.c
@@ -150,12 +150,23 @@ MD5Final(md5byte digest[16], struct MD5Context *ctx) {
 #define MD5STEP(f,w,x,y,z,in,s) \
  (w += f(x,y,z) + in, w = (w<<s | w>>(32-s)) + x)

+#if defined(__clang__) && defined(__has_attribute)
+#if __has_attribute(no_sanitize)
+#define VPX_NO_UNSIGNED_OVERFLOW_CHECK \
+  __attribute__((no_sanitize("unsigned-integer-overflow")))
+#endif
+#endif
+
+#ifndef VPX_NO_UNSIGNED_OVERFLOW_CHECK
+#define VPX_NO_UNSIGNED_OVERFLOW_CHECK
+#endif
+
 /*
 * The core of the MD5 algorithm, this alters an existing MD5 hash to
 * reflect the addition of 16 longwords of new data.  MD5Update blocks
 * the data and converts bytes into longwords for this routine.
 */
-void
+VPX_NO_UNSIGNED_OVERFLOW_CHECK void
 MD5Transform(UWORD32 buf[4], UWORD32 const in[16]) {
  register UWORD32 a, b, c, d;

@@ -238,4 +249,6 @@ MD5Transform(UWORD32 buf[4], UWORD32 const in[16]) {
  buf[3] += d;
 }

+#undef VPX_NO_UNSIGNED_OVERFLOW_CHECK
+
 #endif
--- a/test/acm_random.h
+++ b/test/acm_random.h
@@ -32,6 +32,12 @@ class ACMRandom {
    return (value >> 15) & 0xffff;
  }

+  int16_t Rand9Signed(void) {
+    // Use 9 bits: values between 255 (0x0FF) and -256 (0x100).
+    const uint32_t value = random_.Generate(512);
+    return static_cast<int16_t>(value) - 256;
+  }
+
  uint8_t Rand8(void) {
    const uint32_t value =
        random_.Generate(testing::internal::Random::kMaxRange);
--- a/test/active_map_test.cc
+++ b/test/active_map_test.cc
@@ -85,5 +85,5 @@ TEST_P(ActiveMapTest, Test) {

 VP9_INSTANTIATE_TEST_CASE(ActiveMapTest,
                          ::testing::Values(::libvpx_test::kRealTime),
-                          ::testing::Range(0, 6));
+                          ::testing::Range(0, 9));
 }  // namespace
--- a/test/add_noise_test.cc
+++ b/test/add_noise_test.cc
@@ -0,0 +1,197 @@
+/*
+ *  Copyright (c) 2016 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 <math.h>
+#include "test/clear_system_state.h"
+#include "test/register_state_check.h"
+#include "third_party/googletest/src/include/gtest/gtest.h"
+#include "./vpx_dsp_rtcd.h"
+#include "vpx/vpx_integer.h"
+#include "vpx_mem/vpx_mem.h"
+
+namespace {
+
+// TODO(jimbankoski): make width and height integers not unsigned.
+typedef void (*AddNoiseFunc)(unsigned char *start, char *noise,
+                             char blackclamp[16], char whiteclamp[16],
+                             char bothclamp[16], unsigned int width,
+                             unsigned int height, int pitch);
+
+class AddNoiseTest
+    : public ::testing::TestWithParam<AddNoiseFunc> {
+ public:
+  virtual void TearDown() {
+    libvpx_test::ClearSystemState();
+  }
+  virtual ~AddNoiseTest() {}
+};
+
+double stddev6(char a, char b, char c, char d, char e, char f) {
+  const double n = (a + b + c + d + e + f) / 6.0;
+  const double v = ((a - n) * (a - n) + (b - n) * (b - n) + (c - n) * (c - n) +
+                    (d - n) * (d - n) + (e - n) * (e - n) + (f - n) * (f - n)) /
+                   6.0;
+  return sqrt(v);
+}
+
+// TODO(jimbankoski): The following 2 functions are duplicated in each codec.
+// For now the vp9 one has been copied into the test as is. We should normalize
+// these in vpx_dsp and not have 3 copies of these unless there is different
+// noise we add for each codec.
+
+double gaussian(double sigma, double mu, double x) {
+  return 1 / (sigma * sqrt(2.0 * 3.14159265)) *
+         (exp(-(x - mu) * (x - mu) / (2 * sigma * sigma)));
+}
+
+int setup_noise(int size_noise, char *noise) {
+  char char_dist[300];
+  const int ai = 4;
+  const int qi = 24;
+  const double sigma = ai + .5 + .6 * (63 - qi) / 63.0;
+
+  /* set up a lookup table of 256 entries that matches
+   * a gaussian distribution with sigma determined by q.
+   */
+  int next = 0;
+
+  for (int i = -32; i < 32; i++) {
+    int a_i = (int) (0.5 + 256 * gaussian(sigma, 0, i));
+
+    if (a_i) {
+      for (int j = 0; j < a_i; j++) {
+        char_dist[next + j] = (char)(i);
+      }
+
+      next = next + a_i;
+    }
+  }
+
+  for (; next < 256; next++)
+    char_dist[next] = 0;
+
+  for (int i = 0; i < size_noise; i++) {
+    noise[i] = char_dist[rand() & 0xff];  // NOLINT
+  }
+
+  // Returns the most negative value in distribution.
+  return char_dist[0];
+}
+
+TEST_P(AddNoiseTest, CheckNoiseAdded) {
+  DECLARE_ALIGNED(16, char, blackclamp[16]);
+  DECLARE_ALIGNED(16, char, whiteclamp[16]);
+  DECLARE_ALIGNED(16, char, bothclamp[16]);
+  const int width  = 64;
+  const int height = 64;
+  const int image_size = width * height;
+  char noise[3072];
+
+  const int clamp = setup_noise(3072, noise);
+  for (int i = 0; i < 16; i++) {
+    blackclamp[i] = -clamp;
+    whiteclamp[i] = -clamp;
+    bothclamp[i] = -2 * clamp;
+  }
+
+  uint8_t *const s = reinterpret_cast<uint8_t *>(vpx_calloc(image_size, 1));
+  memset(s, 99, image_size);
+
+  ASM_REGISTER_STATE_CHECK(GetParam()(s, noise, blackclamp, whiteclamp,
+                                      bothclamp, width, height, width));
+
+  // Check to make sure we don't end up having either the same or no added
+  // noise either vertically or horizontally.
+  for (int i = 0; i < image_size - 6 * width - 6; ++i) {
+    const double hd = stddev6(s[i] - 99, s[i + 1] - 99, s[i + 2] - 99,
+                              s[i + 3] - 99, s[i + 4] - 99, s[i + 5] - 99);
+    const double vd = stddev6(s[i] - 99, s[i + width] - 99,
+                              s[i + 2 * width] - 99, s[i + 3 * width] - 99,
+                              s[i + 4 * width] - 99, s[i + 5 * width] - 99);
+
+    EXPECT_NE(hd, 0);
+    EXPECT_NE(vd, 0);
+  }
+
+  // Initialize pixels in the image to 255 and check for roll over.
+  memset(s, 255, image_size);
+
+  ASM_REGISTER_STATE_CHECK(GetParam()(s, noise, blackclamp, whiteclamp,
+                                      bothclamp, width, height, width));
+
+  // Check to make sure don't roll over.
+  for (int i = 0; i < image_size; ++i) {
+    EXPECT_GT((int)s[i], 10) << "i = " << i;
+  }
+
+  // Initialize pixels in the image to 0 and check for roll under.
+  memset(s, 0, image_size);
+
+  ASM_REGISTER_STATE_CHECK(GetParam()(s, noise, blackclamp, whiteclamp,
+                                      bothclamp, width, height, width));
+
+  // Check to make sure don't roll under.
+  for (int i = 0; i < image_size; ++i) {
+    EXPECT_LT((int)s[i], 245) << "i = " << i;
+  }
+
+  vpx_free(s);
+}
+
+TEST_P(AddNoiseTest, CheckCvsAssembly) {
+  DECLARE_ALIGNED(16, char, blackclamp[16]);
+  DECLARE_ALIGNED(16, char, whiteclamp[16]);
+  DECLARE_ALIGNED(16, char, bothclamp[16]);
+  const int width  = 64;
+  const int height = 64;
+  const int image_size = width * height;
+  char noise[3072];
+
+  const int clamp = setup_noise(3072, noise);
+  for (int i = 0; i < 16; i++) {
+    blackclamp[i] = -clamp;
+    whiteclamp[i] = -clamp;
+    bothclamp[i] = -2 * clamp;
+  }
+
+  uint8_t *const s = reinterpret_cast<uint8_t *>(vpx_calloc(image_size, 1));
+  uint8_t *const d = reinterpret_cast<uint8_t *>(vpx_calloc(image_size, 1));
+
+  memset(s, 99, image_size);
+  memset(d, 99, image_size);
+
+  srand(0);
+  ASM_REGISTER_STATE_CHECK(GetParam()(s, noise, blackclamp, whiteclamp,
+                                      bothclamp, width, height, width));
+  srand(0);
+  ASM_REGISTER_STATE_CHECK(vpx_plane_add_noise_c(d, noise, blackclamp,
+                                                 whiteclamp, bothclamp,
+                                                 width, height, width));
+
+  for (int i = 0; i < image_size; ++i) {
+    EXPECT_EQ((int)s[i], (int)d[i]) << "i = " << i;
+  }
+
+  vpx_free(d);
+  vpx_free(s);
+}
+
+INSTANTIATE_TEST_CASE_P(C, AddNoiseTest,
+                        ::testing::Values(vpx_plane_add_noise_c));
+
+#if HAVE_SSE2
+INSTANTIATE_TEST_CASE_P(SSE2, AddNoiseTest,
+                        ::testing::Values(vpx_plane_add_noise_sse2));
+#endif
+
+#if HAVE_MSA
+INSTANTIATE_TEST_CASE_P(MSA, AddNoiseTest,
+                        ::testing::Values(vpx_plane_add_noise_msa));
+#endif
+}  // namespace
--- a/test/altref_test.cc
+++ b/test/altref_test.cc
@@ -14,6 +14,8 @@
 #include "test/util.h"
 namespace {

+#if CONFIG_VP8_ENCODER
+
 // lookahead range: [kLookAheadMin, kLookAheadMax).
 const int kLookAheadMin = 5;
 const int kLookAheadMax = 26;
@@ -63,7 +65,95 @@ TEST_P(AltRefTest, MonotonicTimestamps) {
  EXPECT_GE(altref_count(), 1);
 }

-
 VP8_INSTANTIATE_TEST_CASE(AltRefTest,
                          ::testing::Range(kLookAheadMin, kLookAheadMax));
+
+#endif  // CONFIG_VP8_ENCODER
+
+class AltRefForcedKeyTestLarge
+    : public ::libvpx_test::EncoderTest,
+      public ::libvpx_test::CodecTestWith2Params<libvpx_test::TestMode, int> {
+ protected:
+  AltRefForcedKeyTestLarge()
+      : EncoderTest(GET_PARAM(0)),
+        encoding_mode_(GET_PARAM(1)),
+        cpu_used_(GET_PARAM(2)),
+        forced_kf_frame_num_(1),
+        frame_num_(0) {}
+  virtual ~AltRefForcedKeyTestLarge() {}
+
+  virtual void SetUp() {
+    InitializeConfig();
+    SetMode(encoding_mode_);
+    cfg_.rc_end_usage = VPX_VBR;
+    cfg_.g_threads = 0;
+  }
+
+  virtual void PreEncodeFrameHook(::libvpx_test::VideoSource *video,
+                                  ::libvpx_test::Encoder *encoder) {
+    if (video->frame() == 0) {
+      encoder->Control(VP8E_SET_CPUUSED, cpu_used_);
+      encoder->Control(VP8E_SET_ENABLEAUTOALTREF, 1);
+      // override test default for tile columns if necessary.
+#if CONFIG_VP9_ENCODER
+      if (GET_PARAM(0) == &libvpx_test::kVP9) {
+        encoder->Control(VP9E_SET_TILE_COLUMNS, 6);
+      }
+#endif
+    }
+    frame_flags_ =
+        (video->frame() == forced_kf_frame_num_) ? VPX_EFLAG_FORCE_KF : 0;
+  }
+
+  virtual void FramePktHook(const vpx_codec_cx_pkt_t *pkt) {
+    if (frame_num_ == forced_kf_frame_num_) {
+      ASSERT_TRUE(!!(pkt->data.frame.flags & VPX_FRAME_IS_KEY))
+          << "Frame #" << frame_num_ << " isn't a keyframe!";
+    }
+    ++frame_num_;
+  }
+
+  ::libvpx_test::TestMode encoding_mode_;
+  int cpu_used_;
+  unsigned int forced_kf_frame_num_;
+  unsigned int frame_num_;
+};
+
+TEST_P(AltRefForcedKeyTestLarge, Frame1IsKey) {
+  const vpx_rational timebase = { 1, 30 };
+  const int lag_values[] = { 3, 15, 25, -1 };
+
+  forced_kf_frame_num_ = 1;
+  for (int i = 0; lag_values[i] != -1; ++i) {
+    frame_num_ = 0;
+    cfg_.g_lag_in_frames = lag_values[i];
+    libvpx_test::I420VideoSource video("hantro_collage_w352h288.yuv", 352, 288,
+                                       timebase.den, timebase.num, 0, 30);
+    ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  }
+}
+
+TEST_P(AltRefForcedKeyTestLarge, ForcedFrameIsKey) {
+  const vpx_rational timebase = { 1, 30 };
+  const int lag_values[] = { 3, 15, 25, -1 };
+
+  for (int i = 0; lag_values[i] != -1; ++i) {
+    frame_num_ = 0;
+    forced_kf_frame_num_ = lag_values[i] - 1;
+    cfg_.g_lag_in_frames = lag_values[i];
+    libvpx_test::I420VideoSource video("hantro_collage_w352h288.yuv", 352, 288,
+                                       timebase.den, timebase.num, 0, 30);
+    ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  }
+}
+
+VP8_INSTANTIATE_TEST_CASE(
+    AltRefForcedKeyTestLarge,
+    ::testing::Values(::libvpx_test::kOnePassGood),
+    ::testing::Range(0, 9));
+
+VP9_INSTANTIATE_TEST_CASE(
+    AltRefForcedKeyTestLarge,
+    ::testing::Values(::libvpx_test::kOnePassGood),
+    ::testing::Range(0, 9));
 }  // namespace
--- a/test/avg_test.cc
+++ b/test/avg_test.cc
@@ -55,19 +55,19 @@ class AverageTestBase : public ::testing::Test {
  }

  // Sum Pixels
-  unsigned int ReferenceAverage8x8(const uint8_t* source, int pitch ) {
+  unsigned int ReferenceAverage8x8(const uint8_t* source, int pitch) {
    unsigned int average = 0;
    for (int h = 0; h < 8; ++h)
      for (int w = 0; w < 8; ++w)
-        average += source[h * source_stride_ + w];
+        average += source[h * pitch + w];
    return ((average + 32) >> 6);
  }

-  unsigned int ReferenceAverage4x4(const uint8_t* source, int pitch ) {
+  unsigned int ReferenceAverage4x4(const uint8_t* source, int pitch) {
    unsigned int average = 0;
    for (int h = 0; h < 4; ++h)
      for (int w = 0; w < 4; ++w)
-        average += source[h * source_stride_ + w];
+        average += source[h * pitch + w];
    return ((average + 8) >> 4);
  }

--- a/test/borders_test.cc
+++ b/test/borders_test.cc
@@ -52,7 +52,7 @@ TEST_P(BordersTest, TestEncodeHighBitrate) {
  // extend into the border and test the border condition.
  cfg_.g_lag_in_frames = 25;
  cfg_.rc_2pass_vbr_minsection_pct = 5;
-  cfg_.rc_2pass_vbr_minsection_pct = 2000;
+  cfg_.rc_2pass_vbr_maxsection_pct = 2000;
  cfg_.rc_target_bitrate = 2000;
  cfg_.rc_max_quantizer = 10;

@@ -80,7 +80,4 @@ TEST_P(BordersTest, TestLowBitrate) {

 VP9_INSTANTIATE_TEST_CASE(BordersTest, ::testing::Values(
    ::libvpx_test::kTwoPassGood));
-
-VP10_INSTANTIATE_TEST_CASE(BordersTest, ::testing::Values(
-    ::libvpx_test::kTwoPassGood));
 }  // namespace
--- a/test/byte_alignment_test.cc
+++ b/test/byte_alignment_test.cc
@@ -21,14 +21,14 @@

 namespace {

+#if CONFIG_WEBM_IO
+
 const int kLegacyByteAlignment = 0;
 const int kLegacyYPlaneByteAlignment = 32;
 const int kNumPlanesToCheck = 3;
 const char kVP9TestFile[] = "vp90-2-02-size-lf-1920x1080.webm";
 const char kVP9Md5File[] = "vp90-2-02-size-lf-1920x1080.webm.md5";

-#if CONFIG_WEBM_IO
-
 struct ByteAlignmentTestParam {
  int byte_alignment;
  vpx_codec_err_t expected_value;
--- a/test/codec_factory.h
+++ b/test/codec_factory.h
@@ -13,10 +13,10 @@
 #include "./vpx_config.h"
 #include "vpx/vpx_decoder.h"
 #include "vpx/vpx_encoder.h"
-#if CONFIG_VP8_ENCODER || CONFIG_VP9_ENCODER || CONFIG_VP10_ENCODER
+#if CONFIG_VP8_ENCODER || CONFIG_VP9_ENCODER
 #include "vpx/vp8cx.h"
 #endif
-#if CONFIG_VP8_DECODER || CONFIG_VP9_DECODER || CONFIG_VP10_DECODER
+#if CONFIG_VP8_DECODER || CONFIG_VP9_DECODER
 #include "vpx/vp8dx.h"
 #endif

@@ -233,8 +233,6 @@ class VP9CodecFactory : public CodecFactory {
                                               int usage) const {
 #if CONFIG_VP9_ENCODER
    return vpx_codec_enc_config_default(&vpx_codec_vp9_cx_algo, cfg, usage);
-#elif CONFIG_VP10_ENCODER
-    return vpx_codec_enc_config_default(&vpx_codec_vp10_cx_algo, cfg, usage);
 #else
    return VPX_CODEC_INCAPABLE;
 #endif
@@ -253,96 +251,5 @@ const libvpx_test::VP9CodecFactory kVP9;
 #define VP9_INSTANTIATE_TEST_CASE(test, ...)
 #endif  // CONFIG_VP9

-/*
- * VP10 Codec Definitions
- */
-#if CONFIG_VP10
-class VP10Decoder : public Decoder {
- public:
-  VP10Decoder(vpx_codec_dec_cfg_t cfg, unsigned long deadline)
-      : Decoder(cfg, deadline) {}
-
-  VP10Decoder(vpx_codec_dec_cfg_t cfg, const vpx_codec_flags_t flag,
-              unsigned long deadline)  // NOLINT
-      : Decoder(cfg, flag, deadline) {}
-
- protected:
-  virtual vpx_codec_iface_t* CodecInterface() const {
-#if CONFIG_VP10_DECODER
-    return &vpx_codec_vp10_dx_algo;
-#else
-    return NULL;
-#endif
-  }
-};
-
-class VP10Encoder : public Encoder {
- public:
-  VP10Encoder(vpx_codec_enc_cfg_t cfg, unsigned long deadline,
-              const unsigned long init_flags, TwopassStatsStore *stats)
-      : Encoder(cfg, deadline, init_flags, stats) {}
-
- protected:
-  virtual vpx_codec_iface_t* CodecInterface() const {
-#if CONFIG_VP10_ENCODER
-    return &vpx_codec_vp10_cx_algo;
-#else
-    return NULL;
-#endif
-  }
-};
-
-class VP10CodecFactory : public CodecFactory {
- public:
-  VP10CodecFactory() : CodecFactory() {}
-
-  virtual Decoder* CreateDecoder(vpx_codec_dec_cfg_t cfg,
-                                 unsigned long deadline) const {
-    return CreateDecoder(cfg, 0, deadline);
-  }
-
-  virtual Decoder* CreateDecoder(vpx_codec_dec_cfg_t cfg,
-                                 const vpx_codec_flags_t flags,
-                                 unsigned long deadline) const {  // NOLINT
-#if CONFIG_VP10_DECODER
-    return new VP10Decoder(cfg, flags, deadline);
-#else
-    return NULL;
-#endif
-  }
-
-  virtual Encoder* CreateEncoder(vpx_codec_enc_cfg_t cfg,
-                                 unsigned long deadline,
-                                 const unsigned long init_flags,
-                                 TwopassStatsStore *stats) const {
-#if CONFIG_VP10_ENCODER
-    return new VP10Encoder(cfg, deadline, init_flags, stats);
-#else
-    return NULL;
-#endif
-  }
-
-  virtual vpx_codec_err_t DefaultEncoderConfig(vpx_codec_enc_cfg_t *cfg,
-                                               int usage) const {
-#if CONFIG_VP10_ENCODER
-    return vpx_codec_enc_config_default(&vpx_codec_vp10_cx_algo, cfg, usage);
-#else
-    return VPX_CODEC_INCAPABLE;
-#endif
-  }
-};
-
-const libvpx_test::VP10CodecFactory kVP10;
-
-#define VP10_INSTANTIATE_TEST_CASE(test, ...)\
-  INSTANTIATE_TEST_CASE_P(VP10, test, \
-      ::testing::Combine( \
-          ::testing::Values(static_cast<const libvpx_test::CodecFactory*>( \
-               &libvpx_test::kVP10)), \
-          __VA_ARGS__))
-#else
-#define VP10_INSTANTIATE_TEST_CASE(test, ...)
-#endif  // CONFIG_VP10
-
 }  // namespace libvpx_test
 #endif  // TEST_CODEC_FACTORY_H_
--- a/test/convolve_test.cc
+++ b/test/convolve_test.cc
@@ -69,6 +69,21 @@ struct ConvolveFunctions {

 typedef std::tr1::tuple<int, int, const ConvolveFunctions *> ConvolveParam;

+#define ALL_SIZES(convolve_fn) \
+    make_tuple(4, 4, &convolve_fn),     \
+    make_tuple(8, 4, &convolve_fn),     \
+    make_tuple(4, 8, &convolve_fn),     \
+    make_tuple(8, 8, &convolve_fn),     \
+    make_tuple(16, 8, &convolve_fn),    \
+    make_tuple(8, 16, &convolve_fn),    \
+    make_tuple(16, 16, &convolve_fn),   \
+    make_tuple(32, 16, &convolve_fn),   \
+    make_tuple(16, 32, &convolve_fn),   \
+    make_tuple(32, 32, &convolve_fn),   \
+    make_tuple(64, 32, &convolve_fn),   \
+    make_tuple(32, 64, &convolve_fn),   \
+    make_tuple(64, 64, &convolve_fn)
+
 // Reference 8-tap subpixel filter, slightly modified to fit into this test.
 #define VP9_FILTER_WEIGHT 128
 #define VP9_FILTER_SHIFT 7
@@ -103,7 +118,8 @@ void filter_block2d_8_c(const uint8_t *src_ptr,
  // and filter_max_width          = 16
  //
  uint8_t intermediate_buffer[71 * kMaxDimension];
-  const int intermediate_next_stride = 1 - intermediate_height * output_width;
+  const int intermediate_next_stride =
+      1 - static_cast<int>(intermediate_height * output_width);

  // Horizontal pass (src -> transposed intermediate).
  uint8_t *output_ptr = intermediate_buffer;
@@ -215,7 +231,8 @@ void highbd_filter_block2d_8_c(const uint16_t *src_ptr,
   * and filter_max_width = 16
   */
  uint16_t intermediate_buffer[71 * kMaxDimension];
-  const int intermediate_next_stride = 1 - intermediate_height * output_width;
+  const int intermediate_next_stride =
+      1 - static_cast<int>(intermediate_height * output_width);

  // Horizontal pass (src -> transposed intermediate).
  {
@@ -279,8 +296,7 @@ void highbd_block2d_average_c(uint16_t *src,
                              uint16_t *output_ptr,
                              unsigned int output_stride,
                              unsigned int output_width,
-                              unsigned int output_height,
-                              int bd) {
+                              unsigned int output_height) {
  unsigned int i, j;
  for (i = 0; i < output_height; ++i) {
    for (j = 0; j < output_width; ++j) {
@@ -306,7 +322,7 @@ void highbd_filter_average_block2d_8_c(const uint16_t *src_ptr,
  highbd_filter_block2d_8_c(src_ptr, src_stride, HFilter, VFilter, tmp, 64,
                            output_width, output_height, bd);
  highbd_block2d_average_c(tmp, 64, dst_ptr, dst_stride,
-                           output_width, output_height, bd);
+                           output_width, output_height);
 }
 #endif  // CONFIG_VP9_HIGHBITDEPTH

@@ -1035,20 +1051,6 @@ const ConvolveFunctions convolve8_c(
    wrap_convolve8_horiz_c_8, wrap_convolve8_avg_horiz_c_8,
    wrap_convolve8_vert_c_8, wrap_convolve8_avg_vert_c_8,
    wrap_convolve8_c_8, wrap_convolve8_avg_c_8, 8);
-INSTANTIATE_TEST_CASE_P(C_8, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_c),
-    make_tuple(8, 4, &convolve8_c),
-    make_tuple(4, 8, &convolve8_c),
-    make_tuple(8, 8, &convolve8_c),
-    make_tuple(16, 8, &convolve8_c),
-    make_tuple(8, 16, &convolve8_c),
-    make_tuple(16, 16, &convolve8_c),
-    make_tuple(32, 16, &convolve8_c),
-    make_tuple(16, 32, &convolve8_c),
-    make_tuple(32, 32, &convolve8_c),
-    make_tuple(64, 32, &convolve8_c),
-    make_tuple(32, 64, &convolve8_c),
-    make_tuple(64, 64, &convolve8_c)));
 const ConvolveFunctions convolve10_c(
    wrap_convolve_copy_c_10, wrap_convolve_avg_c_10,
    wrap_convolve8_horiz_c_10, wrap_convolve8_avg_horiz_c_10,
@@ -1057,20 +1059,6 @@ const ConvolveFunctions convolve10_c(
    wrap_convolve8_horiz_c_10, wrap_convolve8_avg_horiz_c_10,
    wrap_convolve8_vert_c_10, wrap_convolve8_avg_vert_c_10,
    wrap_convolve8_c_10, wrap_convolve8_avg_c_10, 10);
-INSTANTIATE_TEST_CASE_P(C_10, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve10_c),
-    make_tuple(8, 4, &convolve10_c),
-    make_tuple(4, 8, &convolve10_c),
-    make_tuple(8, 8, &convolve10_c),
-    make_tuple(16, 8, &convolve10_c),
-    make_tuple(8, 16, &convolve10_c),
-    make_tuple(16, 16, &convolve10_c),
-    make_tuple(32, 16, &convolve10_c),
-    make_tuple(16, 32, &convolve10_c),
-    make_tuple(32, 32, &convolve10_c),
-    make_tuple(64, 32, &convolve10_c),
-    make_tuple(32, 64, &convolve10_c),
-    make_tuple(64, 64, &convolve10_c)));
 const ConvolveFunctions convolve12_c(
    wrap_convolve_copy_c_12, wrap_convolve_avg_c_12,
    wrap_convolve8_horiz_c_12, wrap_convolve8_avg_horiz_c_12,
@@ -1079,23 +1067,13 @@ const ConvolveFunctions convolve12_c(
    wrap_convolve8_horiz_c_12, wrap_convolve8_avg_horiz_c_12,
    wrap_convolve8_vert_c_12, wrap_convolve8_avg_vert_c_12,
    wrap_convolve8_c_12, wrap_convolve8_avg_c_12, 12);
-INSTANTIATE_TEST_CASE_P(C_12, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve12_c),
-    make_tuple(8, 4, &convolve12_c),
-    make_tuple(4, 8, &convolve12_c),
-    make_tuple(8, 8, &convolve12_c),
-    make_tuple(16, 8, &convolve12_c),
-    make_tuple(8, 16, &convolve12_c),
-    make_tuple(16, 16, &convolve12_c),
-    make_tuple(32, 16, &convolve12_c),
-    make_tuple(16, 32, &convolve12_c),
-    make_tuple(32, 32, &convolve12_c),
-    make_tuple(64, 32, &convolve12_c),
-    make_tuple(32, 64, &convolve12_c),
-    make_tuple(64, 64, &convolve12_c)));
+const ConvolveParam kArrayConvolve_c[] = {
+    ALL_SIZES(convolve8_c),
+    ALL_SIZES(convolve10_c),
+    ALL_SIZES(convolve12_c)
+};

 #else
-
 const ConvolveFunctions convolve8_c(
    vpx_convolve_copy_c, vpx_convolve_avg_c,
    vpx_convolve8_horiz_c, vpx_convolve8_avg_horiz_c,
@@ -1104,22 +1082,10 @@ const ConvolveFunctions convolve8_c(
    vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
    vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
    vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
-
-INSTANTIATE_TEST_CASE_P(C, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_c),
-    make_tuple(8, 4, &convolve8_c),
-    make_tuple(4, 8, &convolve8_c),
-    make_tuple(8, 8, &convolve8_c),
-    make_tuple(16, 8, &convolve8_c),
-    make_tuple(8, 16, &convolve8_c),
-    make_tuple(16, 16, &convolve8_c),
-    make_tuple(32, 16, &convolve8_c),
-    make_tuple(16, 32, &convolve8_c),
-    make_tuple(32, 32, &convolve8_c),
-    make_tuple(64, 32, &convolve8_c),
-    make_tuple(32, 64, &convolve8_c),
-    make_tuple(64, 64, &convolve8_c)));
+const ConvolveParam kArrayConvolve_c[] = { ALL_SIZES(convolve8_c) };
 #endif
+INSTANTIATE_TEST_CASE_P(C, ConvolveTest,
+                        ::testing::ValuesIn(kArrayConvolve_c));

 #if HAVE_SSE2 && ARCH_X86_64
 #if CONFIG_VP9_HIGHBITDEPTH
@@ -1159,46 +1125,11 @@ const ConvolveFunctions convolve12_sse2(
    wrap_convolve8_horiz_sse2_12, wrap_convolve8_avg_horiz_sse2_12,
    wrap_convolve8_vert_sse2_12, wrap_convolve8_avg_vert_sse2_12,
    wrap_convolve8_sse2_12, wrap_convolve8_avg_sse2_12, 12);
-INSTANTIATE_TEST_CASE_P(SSE2, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_sse2),
-    make_tuple(8, 4, &convolve8_sse2),
-    make_tuple(4, 8, &convolve8_sse2),
-    make_tuple(8, 8, &convolve8_sse2),
-    make_tuple(16, 8, &convolve8_sse2),
-    make_tuple(8, 16, &convolve8_sse2),
-    make_tuple(16, 16, &convolve8_sse2),
-    make_tuple(32, 16, &convolve8_sse2),
-    make_tuple(16, 32, &convolve8_sse2),
-    make_tuple(32, 32, &convolve8_sse2),
-    make_tuple(64, 32, &convolve8_sse2),
-    make_tuple(32, 64, &convolve8_sse2),
-    make_tuple(64, 64, &convolve8_sse2),
-    make_tuple(4, 4, &convolve10_sse2),
-    make_tuple(8, 4, &convolve10_sse2),
-    make_tuple(4, 8, &convolve10_sse2),
-    make_tuple(8, 8, &convolve10_sse2),
-    make_tuple(16, 8, &convolve10_sse2),
-    make_tuple(8, 16, &convolve10_sse2),
-    make_tuple(16, 16, &convolve10_sse2),
-    make_tuple(32, 16, &convolve10_sse2),
-    make_tuple(16, 32, &convolve10_sse2),
-    make_tuple(32, 32, &convolve10_sse2),
-    make_tuple(64, 32, &convolve10_sse2),
-    make_tuple(32, 64, &convolve10_sse2),
-    make_tuple(64, 64, &convolve10_sse2),
-    make_tuple(4, 4, &convolve12_sse2),
-    make_tuple(8, 4, &convolve12_sse2),
-    make_tuple(4, 8, &convolve12_sse2),
-    make_tuple(8, 8, &convolve12_sse2),
-    make_tuple(16, 8, &convolve12_sse2),
-    make_tuple(8, 16, &convolve12_sse2),
-    make_tuple(16, 16, &convolve12_sse2),
-    make_tuple(32, 16, &convolve12_sse2),
-    make_tuple(16, 32, &convolve12_sse2),
-    make_tuple(32, 32, &convolve12_sse2),
-    make_tuple(64, 32, &convolve12_sse2),
-    make_tuple(32, 64, &convolve12_sse2),
-    make_tuple(64, 64, &convolve12_sse2)));
+const ConvolveParam kArrayConvolve_sse2[] = {
+    ALL_SIZES(convolve8_sse2),
+    ALL_SIZES(convolve10_sse2),
+    ALL_SIZES(convolve12_sse2)
+};
 #else
 const ConvolveFunctions convolve8_sse2(
 #if CONFIG_USE_X86INC
@@ -1213,21 +1144,10 @@ const ConvolveFunctions convolve8_sse2(
    vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
    vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);

-INSTANTIATE_TEST_CASE_P(SSE2, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_sse2),
-    make_tuple(8, 4, &convolve8_sse2),
-    make_tuple(4, 8, &convolve8_sse2),
-    make_tuple(8, 8, &convolve8_sse2),
-    make_tuple(16, 8, &convolve8_sse2),
-    make_tuple(8, 16, &convolve8_sse2),
-    make_tuple(16, 16, &convolve8_sse2),
-    make_tuple(32, 16, &convolve8_sse2),
-    make_tuple(16, 32, &convolve8_sse2),
-    make_tuple(32, 32, &convolve8_sse2),
-    make_tuple(64, 32, &convolve8_sse2),
-    make_tuple(32, 64, &convolve8_sse2),
-    make_tuple(64, 64, &convolve8_sse2)));
+const ConvolveParam kArrayConvolve_sse2[] = { ALL_SIZES(convolve8_sse2) };
 #endif  // CONFIG_VP9_HIGHBITDEPTH
+INSTANTIATE_TEST_CASE_P(SSE2, ConvolveTest,
+                        ::testing::ValuesIn(kArrayConvolve_sse2));
 #endif

 #if HAVE_SSSE3
@@ -1238,22 +1158,11 @@ const ConvolveFunctions convolve8_ssse3(
    vpx_convolve8_ssse3, vpx_convolve8_avg_ssse3,
    vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c,
    vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
-    vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
+    vpx_scaled_2d_ssse3, vpx_scaled_avg_2d_c, 0);

-INSTANTIATE_TEST_CASE_P(SSSE3, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_ssse3),
-    make_tuple(8, 4, &convolve8_ssse3),
-    make_tuple(4, 8, &convolve8_ssse3),
-    make_tuple(8, 8, &convolve8_ssse3),
-    make_tuple(16, 8, &convolve8_ssse3),
-    make_tuple(8, 16, &convolve8_ssse3),
-    make_tuple(16, 16, &convolve8_ssse3),
-    make_tuple(32, 16, &convolve8_ssse3),
-    make_tuple(16, 32, &convolve8_ssse3),
-    make_tuple(32, 32, &convolve8_ssse3),
-    make_tuple(64, 32, &convolve8_ssse3),
-    make_tuple(32, 64, &convolve8_ssse3),
-    make_tuple(64, 64, &convolve8_ssse3)));
+const ConvolveParam kArrayConvolve8_ssse3[] = { ALL_SIZES(convolve8_ssse3) };
+INSTANTIATE_TEST_CASE_P(SSSE3, ConvolveTest,
+                        ::testing::ValuesIn(kArrayConvolve8_ssse3));
 #endif

 #if HAVE_AVX2 && HAVE_SSSE3
@@ -1266,20 +1175,9 @@ const ConvolveFunctions convolve8_avx2(
    vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
    vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);

-INSTANTIATE_TEST_CASE_P(AVX2, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_avx2),
-    make_tuple(8, 4, &convolve8_avx2),
-    make_tuple(4, 8, &convolve8_avx2),
-    make_tuple(8, 8, &convolve8_avx2),
-    make_tuple(8, 16, &convolve8_avx2),
-    make_tuple(16, 8, &convolve8_avx2),
-    make_tuple(16, 16, &convolve8_avx2),
-    make_tuple(32, 16, &convolve8_avx2),
-    make_tuple(16, 32, &convolve8_avx2),
-    make_tuple(32, 32, &convolve8_avx2),
-    make_tuple(64, 32, &convolve8_avx2),
-    make_tuple(32, 64, &convolve8_avx2),
-    make_tuple(64, 64, &convolve8_avx2)));
+const ConvolveParam kArrayConvolve8_avx2[] = { ALL_SIZES(convolve8_avx2) };
+INSTANTIATE_TEST_CASE_P(AVX2, ConvolveTest,
+                        ::testing::ValuesIn(kArrayConvolve8_avx2));
 #endif  // HAVE_AVX2 && HAVE_SSSE3

 #if HAVE_NEON
@@ -1303,20 +1201,9 @@ const ConvolveFunctions convolve8_neon(
    vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
 #endif  // HAVE_NEON_ASM

-INSTANTIATE_TEST_CASE_P(NEON, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_neon),
-    make_tuple(8, 4, &convolve8_neon),
-    make_tuple(4, 8, &convolve8_neon),
-    make_tuple(8, 8, &convolve8_neon),
-    make_tuple(16, 8, &convolve8_neon),
-    make_tuple(8, 16, &convolve8_neon),
-    make_tuple(16, 16, &convolve8_neon),
-    make_tuple(32, 16, &convolve8_neon),
-    make_tuple(16, 32, &convolve8_neon),
-    make_tuple(32, 32, &convolve8_neon),
-    make_tuple(64, 32, &convolve8_neon),
-    make_tuple(32, 64, &convolve8_neon),
-    make_tuple(64, 64, &convolve8_neon)));
+const ConvolveParam kArrayConvolve8_neon[] = { ALL_SIZES(convolve8_neon) };
+INSTANTIATE_TEST_CASE_P(NEON, ConvolveTest,
+                        ::testing::ValuesIn(kArrayConvolve8_neon));
 #endif  // HAVE_NEON

 #if HAVE_DSPR2
@@ -1329,21 +1216,10 @@ const ConvolveFunctions convolve8_dspr2(
    vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
    vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);

-INSTANTIATE_TEST_CASE_P(DSPR2, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_dspr2),
-    make_tuple(8, 4, &convolve8_dspr2),
-    make_tuple(4, 8, &convolve8_dspr2),
-    make_tuple(8, 8, &convolve8_dspr2),
-    make_tuple(16, 8, &convolve8_dspr2),
-    make_tuple(8, 16, &convolve8_dspr2),
-    make_tuple(16, 16, &convolve8_dspr2),
-    make_tuple(32, 16, &convolve8_dspr2),
-    make_tuple(16, 32, &convolve8_dspr2),
-    make_tuple(32, 32, &convolve8_dspr2),
-    make_tuple(64, 32, &convolve8_dspr2),
-    make_tuple(32, 64, &convolve8_dspr2),
-    make_tuple(64, 64, &convolve8_dspr2)));
-#endif
+const ConvolveParam kArrayConvolve8_dspr2[] = { ALL_SIZES(convolve8_dspr2) };
+INSTANTIATE_TEST_CASE_P(DSPR2, ConvolveTest,
+                        ::testing::ValuesIn(kArrayConvolve8_dspr2));
+#endif  // HAVE_DSPR2

 #if HAVE_MSA
 const ConvolveFunctions convolve8_msa(
@@ -1355,19 +1231,8 @@ const ConvolveFunctions convolve8_msa(
    vpx_scaled_vert_c, vpx_scaled_avg_vert_c,
    vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);

-INSTANTIATE_TEST_CASE_P(MSA, ConvolveTest, ::testing::Values(
-    make_tuple(4, 4, &convolve8_msa),
-    make_tuple(8, 4, &convolve8_msa),
-    make_tuple(4, 8, &convolve8_msa),
-    make_tuple(8, 8, &convolve8_msa),
-    make_tuple(16, 8, &convolve8_msa),
-    make_tuple(8, 16, &convolve8_msa),
-    make_tuple(16, 16, &convolve8_msa),
-    make_tuple(32, 16, &convolve8_msa),
-    make_tuple(16, 32, &convolve8_msa),
-    make_tuple(32, 32, &convolve8_msa),
-    make_tuple(64, 32, &convolve8_msa),
-    make_tuple(32, 64, &convolve8_msa),
-    make_tuple(64, 64, &convolve8_msa)));
+const ConvolveParam kArrayConvolve8_msa[] = { ALL_SIZES(convolve8_msa) };
+INSTANTIATE_TEST_CASE_P(MSA, ConvolveTest,
+                        ::testing::ValuesIn(kArrayConvolve8_msa));
 #endif  // HAVE_MSA
 }  // namespace
--- a/test/cpu_speed_test.cc
+++ b/test/cpu_speed_test.cc
@@ -26,7 +26,8 @@ class CpuSpeedTest
      : EncoderTest(GET_PARAM(0)),
        encoding_mode_(GET_PARAM(1)),
        set_cpu_used_(GET_PARAM(2)),
-        min_psnr_(kMaxPSNR) {}
+        min_psnr_(kMaxPSNR),
+        tune_content_(VP9E_CONTENT_DEFAULT) {}
  virtual ~CpuSpeedTest() {}

  virtual void SetUp() {
@@ -49,6 +50,7 @@ class CpuSpeedTest
                                  ::libvpx_test::Encoder *encoder) {
    if (video->frame() == 1) {
      encoder->Control(VP8E_SET_CPUUSED, set_cpu_used_);
+      encoder->Control(VP9E_SET_TUNE_CONTENT, tune_content_);
      if (encoding_mode_ != ::libvpx_test::kRealTime) {
        encoder->Control(VP8E_SET_ENABLEAUTOALTREF, 1);
        encoder->Control(VP8E_SET_ARNR_MAXFRAMES, 7);
@@ -66,6 +68,7 @@ class CpuSpeedTest
  ::libvpx_test::TestMode encoding_mode_;
  int set_cpu_used_;
  double min_psnr_;
+  int tune_content_;
 };

 TEST_P(CpuSpeedTest, TestQ0) {
@@ -74,7 +77,7 @@ TEST_P(CpuSpeedTest, TestQ0) {
  // the encoder to producing lots of big partitions which will likely
  // extend into the border and test the border condition.
  cfg_.rc_2pass_vbr_minsection_pct = 5;
-  cfg_.rc_2pass_vbr_minsection_pct = 2000;
+  cfg_.rc_2pass_vbr_maxsection_pct = 2000;
  cfg_.rc_target_bitrate = 400;
  cfg_.rc_max_quantizer = 0;
  cfg_.rc_min_quantizer = 0;
@@ -92,7 +95,7 @@ TEST_P(CpuSpeedTest, TestScreencastQ0) {
  ::libvpx_test::Y4mVideoSource video("screendata.y4m", 0, 25);
  cfg_.g_timebase = video.timebase();
  cfg_.rc_2pass_vbr_minsection_pct = 5;
-  cfg_.rc_2pass_vbr_minsection_pct = 2000;
+  cfg_.rc_2pass_vbr_maxsection_pct = 2000;
  cfg_.rc_target_bitrate = 400;
  cfg_.rc_max_quantizer = 0;
  cfg_.rc_min_quantizer = 0;
@@ -103,13 +106,28 @@ TEST_P(CpuSpeedTest, TestScreencastQ0) {
  EXPECT_GE(min_psnr_, kMaxPSNR);
 }

+TEST_P(CpuSpeedTest, TestTuneScreen) {
+  ::libvpx_test::Y4mVideoSource video("screendata.y4m", 0, 25);
+  cfg_.g_timebase = video.timebase();
+  cfg_.rc_2pass_vbr_minsection_pct = 5;
+  cfg_.rc_2pass_vbr_minsection_pct = 2000;
+  cfg_.rc_target_bitrate = 2000;
+  cfg_.rc_max_quantizer = 63;
+  cfg_.rc_min_quantizer = 0;
+  tune_content_ = VP9E_CONTENT_SCREEN;
+
+  init_flags_ = VPX_CODEC_USE_PSNR;
+
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+}
+
 TEST_P(CpuSpeedTest, TestEncodeHighBitrate) {
  // Validate that this non multiple of 64 wide clip encodes and decodes
  // without a mismatch when passing in a very low max q.  This pushes
  // the encoder to producing lots of big partitions which will likely
  // extend into the border and test the border condition.
  cfg_.rc_2pass_vbr_minsection_pct = 5;
-  cfg_.rc_2pass_vbr_minsection_pct = 2000;
+  cfg_.rc_2pass_vbr_maxsection_pct = 2000;
  cfg_.rc_target_bitrate = 12000;
  cfg_.rc_max_quantizer = 10;
  cfg_.rc_min_quantizer = 0;
@@ -125,7 +143,7 @@ TEST_P(CpuSpeedTest, TestLowBitrate) {
  // when passing in a very high min q.  This pushes the encoder to producing
  // lots of small partitions which might will test the other condition.
  cfg_.rc_2pass_vbr_minsection_pct = 5;
-  cfg_.rc_2pass_vbr_minsection_pct = 2000;
+  cfg_.rc_2pass_vbr_maxsection_pct = 2000;
  cfg_.rc_target_bitrate = 200;
  cfg_.rc_min_quantizer = 40;

@@ -140,9 +158,4 @@ VP9_INSTANTIATE_TEST_CASE(
    ::testing::Values(::libvpx_test::kTwoPassGood, ::libvpx_test::kOnePassGood,
                      ::libvpx_test::kRealTime),
    ::testing::Range(0, 9));
-
-VP10_INSTANTIATE_TEST_CASE(
-    CpuSpeedTest,
-    ::testing::Values(::libvpx_test::kTwoPassGood, ::libvpx_test::kOnePassGood),
-    ::testing::Range(0, 3));
 }  // namespace
--- a/test/datarate_test.cc
+++ b/test/datarate_test.cc
@@ -90,7 +90,7 @@ class DatarateTestLarge : public ::libvpx_test::EncoderTest,
          << pkt->data.frame.pts;
    }

-    const size_t frame_size_in_bits = pkt->data.frame.sz * 8;
+    const int64_t frame_size_in_bits = pkt->data.frame.sz * 8;

    // Subtract from the buffer the bits associated with a played back frame.
    bits_in_buffer_model_ -= frame_size_in_bits;
@@ -135,7 +135,7 @@ class DatarateTestLarge : public ::libvpx_test::EncoderTest,
  double duration_;
  double file_datarate_;
  double effective_datarate_;
-  size_t bits_in_last_frame_;
+  int64_t bits_in_last_frame_;
  int denoiser_on_;
  int denoiser_offon_test_;
  int denoiser_offon_period_;
@@ -450,7 +450,28 @@ class DatarateTestVP9Large : public ::libvpx_test::EncoderTest,
  int denoiser_offon_period_;
 };

-// Check basic rate targeting,
+// Check basic rate targeting for VBR mode.
+TEST_P(DatarateTestVP9Large, BasicRateTargetingVBR) {
+  cfg_.rc_min_quantizer = 0;
+  cfg_.rc_max_quantizer = 63;
+  cfg_.g_error_resilient = 0;
+  cfg_.rc_end_usage = VPX_VBR;
+  cfg_.g_lag_in_frames = 0;
+
+  ::libvpx_test::I420VideoSource video("hantro_collage_w352h288.yuv", 352, 288,
+                                       30, 1, 0, 300);
+  for (int i = 400; i <= 800; i += 400) {
+    cfg_.rc_target_bitrate = i;
+    ResetModel();
+    ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+    ASSERT_GE(effective_datarate_[0], cfg_.rc_target_bitrate * 0.75)
+        << " The datarate for the file is lower than target by too much!";
+    ASSERT_LE(effective_datarate_[0], cfg_.rc_target_bitrate * 1.25)
+        << " The datarate for the file is greater than target by too much!";
+  }
+}
+
+// Check basic rate targeting for CBR,
 TEST_P(DatarateTestVP9Large, BasicRateTargeting) {
  cfg_.rc_buf_initial_sz = 500;
  cfg_.rc_buf_optimal_sz = 500;
@@ -474,7 +495,7 @@ TEST_P(DatarateTestVP9Large, BasicRateTargeting) {
  }
 }

-// Check basic rate targeting,
+// Check basic rate targeting for CBR.
 TEST_P(DatarateTestVP9Large, BasicRateTargeting444) {
  ::libvpx_test::Y4mVideoSource video("rush_hour_444.y4m", 0, 140);

@@ -519,6 +540,9 @@ TEST_P(DatarateTestVP9Large, ChangingDropFrameThresh) {
  cfg_.rc_end_usage = VPX_CBR;
  cfg_.rc_target_bitrate = 200;
  cfg_.g_lag_in_frames = 0;
+  // TODO(marpan): Investigate datarate target failures with a smaller keyframe
+  // interval (128).
+  cfg_.kf_max_dist = 9999;

  ::libvpx_test::I420VideoSource video("hantro_collage_w352h288.yuv", 352, 288,
                                       30, 1, 0, 140);
@@ -774,10 +798,6 @@ class DatarateOnePassCbrSvc : public ::libvpx_test::EncoderTest,
        svc_params_.max_quantizers[i] = 63;
        svc_params_.min_quantizers[i] = 0;
      }
-      svc_params_.scaling_factor_num[0] = 144;
-      svc_params_.scaling_factor_den[0] = 288;
-      svc_params_.scaling_factor_num[1] = 288;
-      svc_params_.scaling_factor_den[1] = 288;
      encoder->Control(VP9E_SET_SVC, 1);
      encoder->Control(VP9E_SET_SVC_PARAMETERS, &svc_params_);
      encoder->Control(VP8E_SET_CPUUSED, speed_setting_);
@@ -814,8 +834,6 @@ class DatarateOnePassCbrSvc : public ::libvpx_test::EncoderTest,
    if (bits_total_) {
      const double file_size_in_kb = bits_total_ / 1000.;  // bits per kilobit
      duration_ = (last_pts_ + 1) * timebase_;
-      effective_datarate_ = (bits_total_ - bits_in_last_frame_) / 1000.0
-          / (cfg_.rc_buf_initial_sz / 1000.0 + duration_);
      file_datarate_ = file_size_in_kb / duration_;
    }
  }
@@ -839,7 +857,6 @@ class DatarateOnePassCbrSvc : public ::libvpx_test::EncoderTest,
  int64_t bits_total_;
  double duration_;
  double file_datarate_;
-  double effective_datarate_;
  size_t bits_in_last_frame_;
  vpx_svc_extra_cfg_t svc_params_;
  int speed_setting_;
@@ -850,8 +867,7 @@ static void assign_layer_bitrates(vpx_codec_enc_cfg_t *const enc_cfg,
    const vpx_svc_extra_cfg_t *svc_params,
    int spatial_layers,
    int temporal_layers,
-    int temporal_layering_mode,
-    unsigned int total_rate) {
+    int temporal_layering_mode) {
  int sl, spatial_layer_target;
  float total = 0;
  float alloc_ratio[VPX_MAX_LAYERS] = {0};
@@ -885,7 +901,7 @@ static void assign_layer_bitrates(vpx_codec_enc_cfg_t *const enc_cfg,

 // Check basic rate targeting for 1 pass CBR SVC: 2 spatial layers and
 // 3 temporal layers. Run CIF clip with 1 thread.
-TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc) {
+TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc2SpatialLayers) {
  cfg_.rc_buf_initial_sz = 500;
  cfg_.rc_buf_optimal_sz = 500;
  cfg_.rc_buf_sz = 1000;
@@ -905,31 +921,71 @@ TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc) {
  svc_params_.scaling_factor_den[0] = 288;
  svc_params_.scaling_factor_num[1] = 288;
  svc_params_.scaling_factor_den[1] = 288;
-  // TODO(wonkap/marpan): No frame drop for now, we need to implement correct
-  // frame dropping for SVC.
-  cfg_.rc_dropframe_thresh = 0;
+  cfg_.rc_dropframe_thresh = 10;
+  cfg_.kf_max_dist = 9999;
  ::libvpx_test::I420VideoSource video("hantro_collage_w352h288.yuv", 352, 288,
                                       30, 1, 0, 200);
  // TODO(wonkap/marpan): Check that effective_datarate for each layer hits the
-  // layer target_bitrate. Also check if test can pass at lower bitrate (~200k).
-  for (int i = 400; i <= 800; i += 200) {
+  // layer target_bitrate.
+  for (int i = 200; i <= 800; i += 200) {
    cfg_.rc_target_bitrate = i;
    ResetModel();
    assign_layer_bitrates(&cfg_, &svc_params_, cfg_.ss_number_layers,
-        cfg_.ts_number_layers, cfg_.temporal_layering_mode,
-        cfg_.rc_target_bitrate);
+        cfg_.ts_number_layers, cfg_.temporal_layering_mode);
    ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
-    ASSERT_GE(cfg_.rc_target_bitrate, effective_datarate_ * 0.85)
+    ASSERT_GE(cfg_.rc_target_bitrate, file_datarate_ * 0.85)
            << " The datarate for the file exceeds the target by too much!";
    ASSERT_LE(cfg_.rc_target_bitrate, file_datarate_ * 1.15)
        << " The datarate for the file is lower than the target by too much!";
-    EXPECT_EQ(GetMismatchFrames(), (unsigned int) 0);
+    EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
+  }
+}
+
+// Check basic rate targeting for 1 pass CBR SVC: 2 spatial layers and 3
+// temporal layers. Run CIF clip with 1 thread, and few short key frame periods.
+TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc2SpatialLayersSmallKf) {
+  cfg_.rc_buf_initial_sz = 500;
+  cfg_.rc_buf_optimal_sz = 500;
+  cfg_.rc_buf_sz = 1000;
+  cfg_.rc_min_quantizer = 0;
+  cfg_.rc_max_quantizer = 63;
+  cfg_.rc_end_usage = VPX_CBR;
+  cfg_.g_lag_in_frames = 0;
+  cfg_.ss_number_layers = 2;
+  cfg_.ts_number_layers = 3;
+  cfg_.ts_rate_decimator[0] = 4;
+  cfg_.ts_rate_decimator[1] = 2;
+  cfg_.ts_rate_decimator[2] = 1;
+  cfg_.g_error_resilient = 1;
+  cfg_.g_threads = 1;
+  cfg_.temporal_layering_mode = 3;
+  svc_params_.scaling_factor_num[0] = 144;
+  svc_params_.scaling_factor_den[0] = 288;
+  svc_params_.scaling_factor_num[1] = 288;
+  svc_params_.scaling_factor_den[1] = 288;
+  cfg_.rc_dropframe_thresh = 10;
+  ::libvpx_test::I420VideoSource video("hantro_collage_w352h288.yuv", 352, 288,
+                                       30, 1, 0, 200);
+  cfg_.rc_target_bitrate = 400;
+  // For this 3 temporal layer case, pattern repeats every 4 frames, so choose
+  // 4 key neighboring key frame periods (so key frame will land on 0-2-1-2).
+  for (int j = 64; j <= 67; j++) {
+    cfg_.kf_max_dist = j;
+    ResetModel();
+    assign_layer_bitrates(&cfg_, &svc_params_, cfg_.ss_number_layers,
+        cfg_.ts_number_layers, cfg_.temporal_layering_mode);
+    ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+    ASSERT_GE(cfg_.rc_target_bitrate, file_datarate_ * 0.85)
+            << " The datarate for the file exceeds the target by too much!";
+    ASSERT_LE(cfg_.rc_target_bitrate, file_datarate_ * 1.15)
+        << " The datarate for the file is lower than the target by too much!";
+    EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
  }
 }

 // Check basic rate targeting for 1 pass CBR SVC: 2 spatial layers and
 // 3 temporal layers. Run HD clip with 4 threads.
-TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc4threads) {
+TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc2SpatialLayers4threads) {
  cfg_.rc_buf_initial_sz = 500;
  cfg_.rc_buf_optimal_sz = 500;
  cfg_.rc_buf_sz = 1000;
@@ -949,30 +1005,152 @@ TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc4threads) {
  svc_params_.scaling_factor_den[0] = 288;
  svc_params_.scaling_factor_num[1] = 288;
  svc_params_.scaling_factor_den[1] = 288;
-  // TODO(wonkap/marpan): No frame drop for now, we need to implement correct
-  // frame dropping for SVC.
-  cfg_.rc_dropframe_thresh = 0;
+  cfg_.rc_dropframe_thresh = 10;
+  cfg_.kf_max_dist = 9999;
  ::libvpx_test::I420VideoSource video("niklas_1280_720_30.y4m", 1280, 720,
                                       30, 1, 0, 300);
  cfg_.rc_target_bitrate = 800;
  ResetModel();
  assign_layer_bitrates(&cfg_, &svc_params_, cfg_.ss_number_layers,
-      cfg_.ts_number_layers, cfg_.temporal_layering_mode,
-      cfg_.rc_target_bitrate);
+      cfg_.ts_number_layers, cfg_.temporal_layering_mode);
  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
-  ASSERT_GE(cfg_.rc_target_bitrate, effective_datarate_ * 0.85)
+  ASSERT_GE(cfg_.rc_target_bitrate, file_datarate_ * 0.85)
          << " The datarate for the file exceeds the target by too much!";
  ASSERT_LE(cfg_.rc_target_bitrate, file_datarate_ * 1.15)
      << " The datarate for the file is lower than the target by too much!";
-  EXPECT_EQ(GetMismatchFrames(), (unsigned int) 0);
+  EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
+}
+
+// Check basic rate targeting for 1 pass CBR SVC: 3 spatial layers and
+// 3 temporal layers. Run CIF clip with 1 thread.
+TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc3SpatialLayers) {
+  cfg_.rc_buf_initial_sz = 500;
+  cfg_.rc_buf_optimal_sz = 500;
+  cfg_.rc_buf_sz = 1000;
+  cfg_.rc_min_quantizer = 0;
+  cfg_.rc_max_quantizer = 63;
+  cfg_.rc_end_usage = VPX_CBR;
+  cfg_.g_lag_in_frames = 0;
+  cfg_.ss_number_layers = 3;
+  cfg_.ts_number_layers = 3;
+  cfg_.ts_rate_decimator[0] = 4;
+  cfg_.ts_rate_decimator[1] = 2;
+  cfg_.ts_rate_decimator[2] = 1;
+  cfg_.g_error_resilient = 1;
+  cfg_.g_threads = 1;
+  cfg_.temporal_layering_mode = 3;
+  svc_params_.scaling_factor_num[0] = 72;
+  svc_params_.scaling_factor_den[0] = 288;
+  svc_params_.scaling_factor_num[1] = 144;
+  svc_params_.scaling_factor_den[1] = 288;
+  svc_params_.scaling_factor_num[2] = 288;
+  svc_params_.scaling_factor_den[2] = 288;
+  cfg_.rc_dropframe_thresh = 10;
+  cfg_.kf_max_dist = 9999;
+  ::libvpx_test::I420VideoSource video("niklas_1280_720_30.y4m", 1280, 720,
+                                       30, 1, 0, 300);
+  cfg_.rc_target_bitrate = 800;
+  ResetModel();
+  assign_layer_bitrates(&cfg_, &svc_params_, cfg_.ss_number_layers,
+     cfg_.ts_number_layers, cfg_.temporal_layering_mode);
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  ASSERT_GE(cfg_.rc_target_bitrate, file_datarate_ * 0.85)
+          << " The datarate for the file exceeds the target by too much!";
+  ASSERT_LE(cfg_.rc_target_bitrate, file_datarate_ * 1.22)
+      << " The datarate for the file is lower than the target by too much!";
+  EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
+}
+
+// Check basic rate targeting for 1 pass CBR SVC: 3 spatial layers and 3
+// temporal layers. Run CIF clip with 1 thread, and few short key frame periods.
+TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc3SpatialLayersSmallKf) {
+  cfg_.rc_buf_initial_sz = 500;
+  cfg_.rc_buf_optimal_sz = 500;
+  cfg_.rc_buf_sz = 1000;
+  cfg_.rc_min_quantizer = 0;
+  cfg_.rc_max_quantizer = 63;
+  cfg_.rc_end_usage = VPX_CBR;
+  cfg_.g_lag_in_frames = 0;
+  cfg_.ss_number_layers = 3;
+  cfg_.ts_number_layers = 3;
+  cfg_.ts_rate_decimator[0] = 4;
+  cfg_.ts_rate_decimator[1] = 2;
+  cfg_.ts_rate_decimator[2] = 1;
+  cfg_.g_error_resilient = 1;
+  cfg_.g_threads = 1;
+  cfg_.temporal_layering_mode = 3;
+  svc_params_.scaling_factor_num[0] = 72;
+  svc_params_.scaling_factor_den[0] = 288;
+  svc_params_.scaling_factor_num[1] = 144;
+  svc_params_.scaling_factor_den[1] = 288;
+  svc_params_.scaling_factor_num[2] = 288;
+  svc_params_.scaling_factor_den[2] = 288;
+  cfg_.rc_dropframe_thresh = 10;
+  ::libvpx_test::I420VideoSource video("niklas_1280_720_30.y4m", 1280, 720,
+                                       30, 1, 0, 300);
+  cfg_.rc_target_bitrate = 800;
+  // For this 3 temporal layer case, pattern repeats every 4 frames, so choose
+  // 4 key neighboring key frame periods (so key frame will land on 0-2-1-2).
+  for (int j = 32; j <= 35; j++) {
+    cfg_.kf_max_dist = j;
+    ResetModel();
+    assign_layer_bitrates(&cfg_, &svc_params_, cfg_.ss_number_layers,
+       cfg_.ts_number_layers, cfg_.temporal_layering_mode);
+    ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+    ASSERT_GE(cfg_.rc_target_bitrate, file_datarate_ * 0.85)
+            << " The datarate for the file exceeds the target by too much!";
+    ASSERT_LE(cfg_.rc_target_bitrate, file_datarate_ * 1.30)
+        << " The datarate for the file is lower than the target by too much!";
+    EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
+  }
+}
+
+// Check basic rate targeting for 1 pass CBR SVC: 3 spatial layers and
+// 3 temporal layers. Run HD clip with 4 threads.
+TEST_P(DatarateOnePassCbrSvc, OnePassCbrSvc3SpatialLayers4threads) {
+  cfg_.rc_buf_initial_sz = 500;
+  cfg_.rc_buf_optimal_sz = 500;
+  cfg_.rc_buf_sz = 1000;
+  cfg_.rc_min_quantizer = 0;
+  cfg_.rc_max_quantizer = 63;
+  cfg_.rc_end_usage = VPX_CBR;
+  cfg_.g_lag_in_frames = 0;
+  cfg_.ss_number_layers = 3;
+  cfg_.ts_number_layers = 3;
+  cfg_.ts_rate_decimator[0] = 4;
+  cfg_.ts_rate_decimator[1] = 2;
+  cfg_.ts_rate_decimator[2] = 1;
+  cfg_.g_error_resilient = 1;
+  cfg_.g_threads = 4;
+  cfg_.temporal_layering_mode = 3;
+  svc_params_.scaling_factor_num[0] = 72;
+  svc_params_.scaling_factor_den[0] = 288;
+  svc_params_.scaling_factor_num[1] = 144;
+  svc_params_.scaling_factor_den[1] = 288;
+  svc_params_.scaling_factor_num[2] = 288;
+  svc_params_.scaling_factor_den[2] = 288;
+  cfg_.rc_dropframe_thresh = 10;
+  cfg_.kf_max_dist = 9999;
+  ::libvpx_test::I420VideoSource video("niklas_1280_720_30.y4m", 1280, 720,
+                                       30, 1, 0, 300);
+  cfg_.rc_target_bitrate = 800;
+  ResetModel();
+  assign_layer_bitrates(&cfg_, &svc_params_, cfg_.ss_number_layers,
+      cfg_.ts_number_layers, cfg_.temporal_layering_mode);
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  ASSERT_GE(cfg_.rc_target_bitrate, file_datarate_ * 0.85)
+          << " The datarate for the file exceeds the target by too much!";
+  ASSERT_LE(cfg_.rc_target_bitrate, file_datarate_ * 1.22)
+      << " The datarate for the file is lower than the target by too much!";
+  EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
 }

 VP8_INSTANTIATE_TEST_CASE(DatarateTestLarge, ALL_TEST_MODES);
 VP9_INSTANTIATE_TEST_CASE(DatarateTestVP9Large,
                          ::testing::Values(::libvpx_test::kOnePassGood,
                                            ::libvpx_test::kRealTime),
-                          ::testing::Range(2, 7));
+                          ::testing::Range(2, 9));
 VP9_INSTANTIATE_TEST_CASE(DatarateOnePassCbrSvc,
                          ::testing::Values(::libvpx_test::kRealTime),
-                          ::testing::Range(5, 8));
+                          ::testing::Range(5, 9));
 }  // namespace
--- a/test/dct16x16_test.cc
+++ b/test/dct16x16_test.cc
@@ -276,12 +276,12 @@ void idct16x16_12(const tran_low_t *in, uint8_t *out, int stride) {
 }

 void idct16x16_10_ref(const tran_low_t *in, uint8_t *out, int stride,
-                      int tx_type) {
+                      int /*tx_type*/) {
  idct16x16_10(in, out, stride);
 }

 void idct16x16_12_ref(const tran_low_t *in, uint8_t *out, int stride,
-                      int tx_type) {
+                      int /*tx_type*/) {
  idct16x16_12(in, out, stride);
 }

@@ -293,6 +293,7 @@ void iht16x16_12(const tran_low_t *in, uint8_t *out, int stride, int tx_type) {
  vp9_highbd_iht16x16_256_add_c(in, out, stride, tx_type, 12);
 }

+#if HAVE_SSE2
 void idct16x16_10_add_10_c(const tran_low_t *in, uint8_t *out, int stride) {
  vpx_highbd_idct16x16_10_add_c(in, out, stride, 10);
 }
@@ -301,7 +302,6 @@ void idct16x16_10_add_12_c(const tran_low_t *in, uint8_t *out, int stride) {
  vpx_highbd_idct16x16_10_add_c(in, out, stride, 12);
 }

-#if HAVE_SSE2
 void idct16x16_256_add_10_sse2(const tran_low_t *in, uint8_t *out, int stride) {
  vpx_highbd_idct16x16_256_add_sse2(in, out, stride, 10);
 }
@@ -373,10 +373,10 @@ class Trans16x16TestBase {

      for (int j = 0; j < kNumCoeffs; ++j) {
 #if CONFIG_VP9_HIGHBITDEPTH
-        const uint32_t diff =
+        const int32_t diff =
            bit_depth_ == VPX_BITS_8 ?  dst[j] - src[j] : dst16[j] - src16[j];
 #else
-        const uint32_t diff = dst[j] - src[j];
+        const int32_t diff = dst[j] - src[j];
 #endif
        const uint32_t error = diff * diff;
        if (max_error < error)
@@ -778,7 +778,7 @@ class InvTrans16x16DCT
  virtual void TearDown() { libvpx_test::ClearSystemState(); }

 protected:
-  void RunFwdTxfm(int16_t *in, tran_low_t *out, int stride) {}
+  void RunFwdTxfm(int16_t * /*in*/, tran_low_t * /*out*/, int /*stride*/) {}
  void RunInvTxfm(tran_low_t *out, uint8_t *dst, int stride) {
    inv_txfm_(out, dst, stride);
  }
@@ -792,6 +792,67 @@ TEST_P(InvTrans16x16DCT, CompareReference) {
  CompareInvReference(ref_txfm_, thresh_);
 }

+class PartialTrans16x16Test
+    : public ::testing::TestWithParam<
+          std::tr1::tuple<FdctFunc, vpx_bit_depth_t> > {
+ public:
+  virtual ~PartialTrans16x16Test() {}
+  virtual void SetUp() {
+    fwd_txfm_ = GET_PARAM(0);
+    bit_depth_ = GET_PARAM(1);
+  }
+
+  virtual void TearDown() { libvpx_test::ClearSystemState(); }
+
+ protected:
+  vpx_bit_depth_t bit_depth_;
+  FdctFunc fwd_txfm_;
+};
+
+TEST_P(PartialTrans16x16Test, Extremes) {
+#if CONFIG_VP9_HIGHBITDEPTH
+  const int16_t maxval =
+      static_cast<int16_t>(clip_pixel_highbd(1 << 30, bit_depth_));
+#else
+  const int16_t maxval = 255;
+#endif
+  const int minval = -maxval;
+  DECLARE_ALIGNED(16, int16_t, input[kNumCoeffs]);
+  DECLARE_ALIGNED(16, tran_low_t, output[kNumCoeffs]);
+
+  for (int i = 0; i < kNumCoeffs; ++i) input[i] = maxval;
+  output[0] = 0;
+  ASM_REGISTER_STATE_CHECK(fwd_txfm_(input, output, 16));
+  EXPECT_EQ((maxval * kNumCoeffs) >> 1, output[0]);
+
+  for (int i = 0; i < kNumCoeffs; ++i) input[i] = minval;
+  output[0] = 0;
+  ASM_REGISTER_STATE_CHECK(fwd_txfm_(input, output, 16));
+  EXPECT_EQ((minval * kNumCoeffs) >> 1, output[0]);
+}
+
+TEST_P(PartialTrans16x16Test, Random) {
+#if CONFIG_VP9_HIGHBITDEPTH
+  const int16_t maxval =
+      static_cast<int16_t>(clip_pixel_highbd(1 << 30, bit_depth_));
+#else
+  const int16_t maxval = 255;
+#endif
+  DECLARE_ALIGNED(16, int16_t, input[kNumCoeffs]);
+  DECLARE_ALIGNED(16, tran_low_t, output[kNumCoeffs]);
+  ACMRandom rnd(ACMRandom::DeterministicSeed());
+
+  int sum = 0;
+  for (int i = 0; i < kNumCoeffs; ++i) {
+    const int val = (i & 1) ? -rnd(maxval + 1) : rnd(maxval + 1);
+    input[i] = val;
+    sum += val;
+  }
+  output[0] = 0;
+  ASM_REGISTER_STATE_CHECK(fwd_txfm_(input, output, 16));
+  EXPECT_EQ(sum >> 1, output[0]);
+}
+
 using std::tr1::make_tuple;

 #if CONFIG_VP9_HIGHBITDEPTH
@@ -824,6 +885,11 @@ INSTANTIATE_TEST_CASE_P(
        make_tuple(&vp9_fht16x16_c, &vp9_iht16x16_256_add_c, 1, VPX_BITS_8),
        make_tuple(&vp9_fht16x16_c, &vp9_iht16x16_256_add_c, 2, VPX_BITS_8),
        make_tuple(&vp9_fht16x16_c, &vp9_iht16x16_256_add_c, 3, VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(
+    C, PartialTrans16x16Test,
+    ::testing::Values(make_tuple(&vpx_highbd_fdct16x16_1_c, VPX_BITS_8),
+                      make_tuple(&vpx_highbd_fdct16x16_1_c, VPX_BITS_10),
+                      make_tuple(&vpx_highbd_fdct16x16_1_c, VPX_BITS_12)));
 #else
 INSTANTIATE_TEST_CASE_P(
    C, Trans16x16HT,
@@ -832,6 +898,9 @@ INSTANTIATE_TEST_CASE_P(
        make_tuple(&vp9_fht16x16_c, &vp9_iht16x16_256_add_c, 1, VPX_BITS_8),
        make_tuple(&vp9_fht16x16_c, &vp9_iht16x16_256_add_c, 2, VPX_BITS_8),
        make_tuple(&vp9_fht16x16_c, &vp9_iht16x16_256_add_c, 3, VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(C, PartialTrans16x16Test,
+                        ::testing::Values(make_tuple(&vpx_fdct16x16_1_c,
+                                                     VPX_BITS_8)));
 #endif  // CONFIG_VP9_HIGHBITDEPTH

 #if HAVE_NEON_ASM && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
@@ -859,6 +928,9 @@ INSTANTIATE_TEST_CASE_P(
                   VPX_BITS_8),
        make_tuple(&vp9_fht16x16_sse2, &vp9_iht16x16_256_add_sse2, 3,
                   VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(SSE2, PartialTrans16x16Test,
+                        ::testing::Values(make_tuple(&vpx_fdct16x16_1_sse2,
+                                                     VPX_BITS_8)));
 #endif  // HAVE_SSE2 && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE

 #if HAVE_SSE2 && CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
@@ -896,6 +968,9 @@ INSTANTIATE_TEST_CASE_P(
                   &idct16x16_10_add_12_sse2, 3167, VPX_BITS_12),
        make_tuple(&idct16x16_12,
                   &idct16x16_256_add_12_sse2, 3167, VPX_BITS_12)));
+INSTANTIATE_TEST_CASE_P(SSE2, PartialTrans16x16Test,
+                        ::testing::Values(make_tuple(&vpx_fdct16x16_1_sse2,
+                                                     VPX_BITS_8)));
 #endif  // HAVE_SSE2 && CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE

 #if HAVE_MSA && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
@@ -912,5 +987,8 @@ INSTANTIATE_TEST_CASE_P(
        make_tuple(&vp9_fht16x16_msa, &vp9_iht16x16_256_add_msa, 2, VPX_BITS_8),
        make_tuple(&vp9_fht16x16_msa, &vp9_iht16x16_256_add_msa, 3,
                   VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(MSA, PartialTrans16x16Test,
+                        ::testing::Values(make_tuple(&vpx_fdct16x16_1_msa,
+                                                     VPX_BITS_8)));
 #endif  // HAVE_MSA && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
 }  // namespace
--- a/test/dct32x32_test.cc
+++ b/test/dct32x32_test.cc
@@ -81,10 +81,6 @@ typedef std::tr1::tuple<FwdTxfmFunc, InvTxfmFunc, int, vpx_bit_depth_t>
    Trans32x32Param;

 #if CONFIG_VP9_HIGHBITDEPTH
-void idct32x32_8(const tran_low_t *in, uint8_t *out, int stride) {
-  vpx_highbd_idct32x32_1024_add_c(in, out, stride, 8);
-}
-
 void idct32x32_10(const tran_low_t *in, uint8_t *out, int stride) {
  vpx_highbd_idct32x32_1024_add_c(in, out, stride, 10);
 }
@@ -158,10 +154,10 @@ TEST_P(Trans32x32Test, AccuracyCheck) {

    for (int j = 0; j < kNumCoeffs; ++j) {
 #if CONFIG_VP9_HIGHBITDEPTH
-      const uint32_t diff =
+      const int32_t diff =
          bit_depth_ == VPX_BITS_8 ? dst[j] - src[j] : dst16[j] - src16[j];
 #else
-      const uint32_t diff = dst[j] - src[j];
+      const int32_t diff = dst[j] - src[j];
 #endif
      const uint32_t error = diff * diff;
      if (max_error < error)
@@ -309,6 +305,67 @@ TEST_P(Trans32x32Test, InverseAccuracy) {
  }
 }

+class PartialTrans32x32Test
+    : public ::testing::TestWithParam<
+          std::tr1::tuple<FwdTxfmFunc, vpx_bit_depth_t> > {
+ public:
+  virtual ~PartialTrans32x32Test() {}
+  virtual void SetUp() {
+    fwd_txfm_ = GET_PARAM(0);
+    bit_depth_ = GET_PARAM(1);
+  }
+
+  virtual void TearDown() { libvpx_test::ClearSystemState(); }
+
+ protected:
+  vpx_bit_depth_t bit_depth_;
+  FwdTxfmFunc fwd_txfm_;
+};
+
+TEST_P(PartialTrans32x32Test, Extremes) {
+#if CONFIG_VP9_HIGHBITDEPTH
+  const int16_t maxval =
+      static_cast<int16_t>(clip_pixel_highbd(1 << 30, bit_depth_));
+#else
+  const int16_t maxval = 255;
+#endif
+  const int minval = -maxval;
+  DECLARE_ALIGNED(16, int16_t, input[kNumCoeffs]);
+  DECLARE_ALIGNED(16, tran_low_t, output[kNumCoeffs]);
+
+  for (int i = 0; i < kNumCoeffs; ++i) input[i] = maxval;
+  output[0] = 0;
+  ASM_REGISTER_STATE_CHECK(fwd_txfm_(input, output, 32));
+  EXPECT_EQ((maxval * kNumCoeffs) >> 3, output[0]);
+
+  for (int i = 0; i < kNumCoeffs; ++i) input[i] = minval;
+  output[0] = 0;
+  ASM_REGISTER_STATE_CHECK(fwd_txfm_(input, output, 32));
+  EXPECT_EQ((minval * kNumCoeffs) >> 3, output[0]);
+}
+
+TEST_P(PartialTrans32x32Test, Random) {
+#if CONFIG_VP9_HIGHBITDEPTH
+  const int16_t maxval =
+      static_cast<int16_t>(clip_pixel_highbd(1 << 30, bit_depth_));
+#else
+  const int16_t maxval = 255;
+#endif
+  DECLARE_ALIGNED(16, int16_t, input[kNumCoeffs]);
+  DECLARE_ALIGNED(16, tran_low_t, output[kNumCoeffs]);
+  ACMRandom rnd(ACMRandom::DeterministicSeed());
+
+  int sum = 0;
+  for (int i = 0; i < kNumCoeffs; ++i) {
+    const int val = (i & 1) ? -rnd(maxval + 1) : rnd(maxval + 1);
+    input[i] = val;
+    sum += val;
+  }
+  output[0] = 0;
+  ASM_REGISTER_STATE_CHECK(fwd_txfm_(input, output, 32));
+  EXPECT_EQ(sum >> 3, output[0]);
+}
+
 using std::tr1::make_tuple;

 #if CONFIG_VP9_HIGHBITDEPTH
@@ -327,6 +384,11 @@ INSTANTIATE_TEST_CASE_P(
                   &vpx_idct32x32_1024_add_c, 0, VPX_BITS_8),
        make_tuple(&vpx_fdct32x32_rd_c,
                   &vpx_idct32x32_1024_add_c, 1, VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(
+    C, PartialTrans32x32Test,
+    ::testing::Values(make_tuple(&vpx_highbd_fdct32x32_1_c, VPX_BITS_8),
+                      make_tuple(&vpx_highbd_fdct32x32_1_c, VPX_BITS_10),
+                      make_tuple(&vpx_highbd_fdct32x32_1_c, VPX_BITS_12)));
 #else
 INSTANTIATE_TEST_CASE_P(
    C, Trans32x32Test,
@@ -335,9 +397,12 @@ INSTANTIATE_TEST_CASE_P(
                   &vpx_idct32x32_1024_add_c, 0, VPX_BITS_8),
        make_tuple(&vpx_fdct32x32_rd_c,
                   &vpx_idct32x32_1024_add_c, 1, VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(C, PartialTrans32x32Test,
+                        ::testing::Values(make_tuple(&vpx_fdct32x32_1_c,
+                                                     VPX_BITS_8)));
 #endif  // CONFIG_VP9_HIGHBITDEPTH

-#if HAVE_NEON_ASM && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
+#if HAVE_NEON && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
 INSTANTIATE_TEST_CASE_P(
    NEON, Trans32x32Test,
    ::testing::Values(
@@ -345,7 +410,7 @@ INSTANTIATE_TEST_CASE_P(
                   &vpx_idct32x32_1024_add_neon, 0, VPX_BITS_8),
        make_tuple(&vpx_fdct32x32_rd_c,
                   &vpx_idct32x32_1024_add_neon, 1, VPX_BITS_8)));
-#endif  // HAVE_NEON_ASM && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
+#endif  // HAVE_NEON && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE

 #if HAVE_SSE2 && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
 INSTANTIATE_TEST_CASE_P(
@@ -355,6 +420,9 @@ INSTANTIATE_TEST_CASE_P(
                   &vpx_idct32x32_1024_add_sse2, 0, VPX_BITS_8),
        make_tuple(&vpx_fdct32x32_rd_sse2,
                   &vpx_idct32x32_1024_add_sse2, 1, VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(SSE2, PartialTrans32x32Test,
+                        ::testing::Values(make_tuple(&vpx_fdct32x32_1_sse2,
+                                                     VPX_BITS_8)));
 #endif  // HAVE_SSE2 && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE

 #if HAVE_SSE2 && CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
@@ -371,6 +439,9 @@ INSTANTIATE_TEST_CASE_P(
                   VPX_BITS_8),
        make_tuple(&vpx_fdct32x32_rd_sse2, &vpx_idct32x32_1024_add_c, 1,
                   VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(SSE2, PartialTrans32x32Test,
+                        ::testing::Values(make_tuple(&vpx_fdct32x32_1_sse2,
+                                                     VPX_BITS_8)));
 #endif  // HAVE_SSE2 && CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE

 #if HAVE_AVX2 && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
@@ -391,5 +462,8 @@ INSTANTIATE_TEST_CASE_P(
                   &vpx_idct32x32_1024_add_msa, 0, VPX_BITS_8),
        make_tuple(&vpx_fdct32x32_rd_msa,
                   &vpx_idct32x32_1024_add_msa, 1, VPX_BITS_8)));
+INSTANTIATE_TEST_CASE_P(MSA, PartialTrans32x32Test,
+                        ::testing::Values(make_tuple(&vpx_fdct32x32_1_msa,
+                                                     VPX_BITS_8)));
 #endif  // HAVE_MSA && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
 }  // namespace
--- a/test/decode_api_test.cc
+++ b/test/decode_api_test.cc
@@ -26,9 +26,6 @@ TEST(DecodeAPI, InvalidParams) {
 #endif
 #if CONFIG_VP9_DECODER
    &vpx_codec_vp9_dx_algo,
-#endif
-#if CONFIG_VP10_DECODER
-    &vpx_codec_vp10_dx_algo,
 #endif
  };
  uint8_t buf[1] = {0};
@@ -146,6 +143,40 @@ TEST(DecodeAPI, Vp9InvalidDecode) {
  TestVp9Controls(&dec);
  EXPECT_EQ(VPX_CODEC_OK, vpx_codec_destroy(&dec));
 }
+
+TEST(DecodeAPI, Vp9PeekSI) {
+  const vpx_codec_iface_t *const codec = &vpx_codec_vp9_dx_algo;
+  // The first 9 bytes are valid and the rest of the bytes are made up. Until
+  // size 10, this should return VPX_CODEC_UNSUP_BITSTREAM and after that it
+  // should return VPX_CODEC_CORRUPT_FRAME.
+  const uint8_t data[32] = {
+    0x85, 0xa4, 0xc1, 0xa1, 0x38, 0x81, 0xa3, 0x49,
+    0x83, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+    0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+    0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+  };
+
+  for (uint32_t data_sz = 1; data_sz <= 32; ++data_sz) {
+    // Verify behavior of vpx_codec_decode. vpx_codec_decode doesn't even get
+    // to decoder_peek_si_internal on frames of size < 8.
+    if (data_sz >= 8) {
+      vpx_codec_ctx_t dec;
+      EXPECT_EQ(VPX_CODEC_OK, vpx_codec_dec_init(&dec, codec, NULL, 0));
+      EXPECT_EQ((data_sz < 10) ?
+                    VPX_CODEC_UNSUP_BITSTREAM : VPX_CODEC_CORRUPT_FRAME,
+                vpx_codec_decode(&dec, data, data_sz, NULL, 0));
+      vpx_codec_iter_t iter = NULL;
+      EXPECT_EQ(NULL, vpx_codec_get_frame(&dec, &iter));
+      EXPECT_EQ(VPX_CODEC_OK, vpx_codec_destroy(&dec));
+    }
+
+    // Verify behavior of vpx_codec_peek_stream_info.
+    vpx_codec_stream_info_t si;
+    si.sz = sizeof(si);
+    EXPECT_EQ((data_sz < 10) ? VPX_CODEC_UNSUP_BITSTREAM : VPX_CODEC_OK,
+              vpx_codec_peek_stream_info(codec, data, data_sz, &si));
+  }
+}
 #endif  // CONFIG_VP9_DECODER

 }  // namespace
--- a/test/encode_api_test.cc
+++ b/test/encode_api_test.cc
@@ -0,0 +1,65 @@
+/*
+ *  Copyright (c) 2016 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 "third_party/googletest/src/include/gtest/gtest.h"
+
+#include "./vpx_config.h"
+#include "vpx/vp8cx.h"
+#include "vpx/vpx_encoder.h"
+
+namespace {
+
+#define NELEMENTS(x) static_cast<int>(sizeof(x) / sizeof(x[0]))
+
+TEST(EncodeAPI, InvalidParams) {
+  static const vpx_codec_iface_t *kCodecs[] = {
+#if CONFIG_VP8_ENCODER
+    &vpx_codec_vp8_cx_algo,
+#endif
+#if CONFIG_VP9_ENCODER
+    &vpx_codec_vp9_cx_algo,
+#endif
+  };
+  uint8_t buf[1] = {0};
+  vpx_image_t img;
+  vpx_codec_ctx_t enc;
+  vpx_codec_enc_cfg_t cfg;
+
+  EXPECT_EQ(&img, vpx_img_wrap(&img, VPX_IMG_FMT_I420, 1, 1, 1, buf));
+
+  EXPECT_EQ(VPX_CODEC_INVALID_PARAM, vpx_codec_enc_init(NULL, NULL, NULL, 0));
+  EXPECT_EQ(VPX_CODEC_INVALID_PARAM, vpx_codec_enc_init(&enc, NULL, NULL, 0));
+  EXPECT_EQ(VPX_CODEC_INVALID_PARAM, vpx_codec_encode(NULL, NULL, 0, 0, 0, 0));
+  EXPECT_EQ(VPX_CODEC_INVALID_PARAM, vpx_codec_encode(NULL, &img, 0, 0, 0, 0));
+  EXPECT_EQ(VPX_CODEC_INVALID_PARAM, vpx_codec_destroy(NULL));
+  EXPECT_EQ(VPX_CODEC_INVALID_PARAM,
+            vpx_codec_enc_config_default(NULL, NULL, 0));
+  EXPECT_EQ(VPX_CODEC_INVALID_PARAM,
+            vpx_codec_enc_config_default(NULL, &cfg, 0));
+  EXPECT_TRUE(vpx_codec_error(NULL) != NULL);
+
+  for (int i = 0; i < NELEMENTS(kCodecs); ++i) {
+    SCOPED_TRACE(vpx_codec_iface_name(kCodecs[i]));
+    EXPECT_EQ(VPX_CODEC_INVALID_PARAM,
+              vpx_codec_enc_init(NULL, kCodecs[i], NULL, 0));
+    EXPECT_EQ(VPX_CODEC_INVALID_PARAM,
+              vpx_codec_enc_init(&enc, kCodecs[i], NULL, 0));
+    EXPECT_EQ(VPX_CODEC_INVALID_PARAM,
+              vpx_codec_enc_config_default(kCodecs[i], &cfg, 1));
+
+    EXPECT_EQ(VPX_CODEC_OK, vpx_codec_enc_config_default(kCodecs[i], &cfg, 0));
+    EXPECT_EQ(VPX_CODEC_OK, vpx_codec_enc_init(&enc, kCodecs[i], &cfg, 0));
+    EXPECT_EQ(VPX_CODEC_OK, vpx_codec_encode(&enc, NULL, 0, 0, 0, 0));
+
+    EXPECT_EQ(VPX_CODEC_OK, vpx_codec_destroy(&enc));
+  }
+}
+
+}  // namespace
--- a/test/encode_test_driver.cc
+++ b/test/encode_test_driver.cc
@@ -42,15 +42,6 @@ void Encoder::InitEncoder(VideoSource *video) {
                               log2_tile_columns);
      ASSERT_EQ(VPX_CODEC_OK, res) << EncoderError();
    } else
-#endif
-#if CONFIG_VP10_ENCODER
-    if (CodecInterface() == &vpx_codec_vp10_cx_algo) {
-      // Default to 1 tile column for VP10.
-      const int log2_tile_columns = 0;
-      res = vpx_codec_control_(&encoder_, VP9E_SET_TILE_COLUMNS,
-                               log2_tile_columns);
-      ASSERT_EQ(VPX_CODEC_OK, res) << EncoderError();
-    } else
 #endif
    {
 #if CONFIG_VP8_ENCODER
--- a/test/encode_test_driver.h
+++ b/test/encode_test_driver.h
@@ -16,7 +16,7 @@
 #include "third_party/googletest/src/include/gtest/gtest.h"

 #include "./vpx_config.h"
-#if CONFIG_VP8_ENCODER || CONFIG_VP9_ENCODER || CONFIG_VP10_ENCODER
+#if CONFIG_VP8_ENCODER || CONFIG_VP9_ENCODER
 #include "vpx/vp8cx.h"
 #endif
 #include "vpx/vpx_encoder.h"
@@ -143,7 +143,7 @@ class Encoder {
    const vpx_codec_err_t res = vpx_codec_control_(&encoder_, ctrl_id, arg);
    ASSERT_EQ(VPX_CODEC_OK, res) << EncoderError();
  }
-#if CONFIG_VP8_ENCODER || CONFIG_VP9_ENCODER || CONFIG_VP10_ENCODER
+#if CONFIG_VP8_ENCODER || CONFIG_VP9_ENCODER
  void Control(int ctrl_id, vpx_active_map_t *arg) {
    const vpx_codec_err_t res = vpx_codec_control_(&encoder_, ctrl_id, arg);
    ASSERT_EQ(VPX_CODEC_OK, res) << EncoderError();
--- a/test/error_resilience_test.cc
+++ b/test/error_resilience_test.cc
@@ -100,7 +100,7 @@ class ErrorResilienceTestLarge : public ::libvpx_test::EncoderTest,
  }

  virtual void PreEncodeFrameHook(libvpx_test::VideoSource *video,
-                                  ::libvpx_test::Encoder *encoder) {
+                                  ::libvpx_test::Encoder * /*encoder*/) {
    frame_flags_ &= ~(VP8_EFLAG_NO_UPD_LAST |
                      VP8_EFLAG_NO_UPD_GF |
                      VP8_EFLAG_NO_UPD_ARF);
@@ -596,7 +596,4 @@ VP8_INSTANTIATE_TEST_CASE(ErrorResilienceTestLargeCodecControls,
                          ONE_PASS_TEST_MODES);
 VP9_INSTANTIATE_TEST_CASE(ErrorResilienceTestLarge, ONE_PASS_TEST_MODES,
                          ::testing::Values(true));
-// SVC-related tests don't run for VP10 since SVC is not supported.
-VP10_INSTANTIATE_TEST_CASE(ErrorResilienceTestLarge, ONE_PASS_TEST_MODES,
-                           ::testing::Values(false));
 }  // namespace
--- a/test/external_frame_buffer_test.cc
+++ b/test/external_frame_buffer_test.cc
@@ -24,7 +24,6 @@
 namespace {

 const int kVideoNameParam = 1;
-const char kVP9TestFile[] = "vp90-2-02-size-lf-1920x1080.webm";

 struct ExternalFrameBuffer {
  uint8_t *data;
@@ -155,6 +154,8 @@ class ExternalFrameBufferList {
  ExternalFrameBuffer *ext_fb_list_;
 };

+#if CONFIG_WEBM_IO
+
 // Callback used by libvpx to request the application to return a frame
 // buffer of at least |min_size| in bytes.
 int get_vp9_frame_buffer(void *user_priv, size_t min_size,
@@ -197,6 +198,8 @@ int do_not_release_vp9_frame_buffer(void *user_priv,
  return 0;
 }

+#endif  // CONFIG_WEBM_IO
+
 // Class for testing passing in external frame buffers to libvpx.
 class ExternalFrameBufferMD5Test
    : public ::libvpx_test::DecoderTest,
@@ -278,6 +281,8 @@ class ExternalFrameBufferMD5Test
 };

 #if CONFIG_WEBM_IO
+const char kVP9TestFile[] = "vp90-2-02-size-lf-1920x1080.webm";
+
 // Class for testing passing in external frame buffers to libvpx.
 class ExternalFrameBufferTest : public ::testing::Test {
 protected:
--- a/test/fdct4x4_test.cc
+++ b/test/fdct4x4_test.cc
@@ -40,7 +40,7 @@ typedef std::tr1::tuple<FdctFunc, IdctFunc, int, vpx_bit_depth_t> Dct4x4Param;
 typedef std::tr1::tuple<FhtFunc, IhtFunc, int, vpx_bit_depth_t> Ht4x4Param;

 void fdct4x4_ref(const int16_t *in, tran_low_t *out, int stride,
-                 int tx_type) {
+                 int /*tx_type*/) {
  vpx_fdct4x4_c(in, out, stride);
 }

@@ -49,7 +49,7 @@ void fht4x4_ref(const int16_t *in, tran_low_t *out, int stride, int tx_type) {
 }

 void fwht4x4_ref(const int16_t *in, tran_low_t *out, int stride,
-                 int tx_type) {
+                 int /*tx_type*/) {
  vp9_fwht4x4_c(in, out, stride);
 }

@@ -141,11 +141,11 @@ class Trans4x4TestBase {

      for (int j = 0; j < kNumCoeffs; ++j) {
 #if CONFIG_VP9_HIGHBITDEPTH
-        const uint32_t diff =
+        const int diff =
            bit_depth_ == VPX_BITS_8 ? dst[j] - src[j] : dst16[j] - src16[j];
 #else
        ASSERT_EQ(VPX_BITS_8, bit_depth_);
-        const uint32_t diff = dst[j] - src[j];
+        const int diff = dst[j] - src[j];
 #endif
        const uint32_t error = diff * diff;
        if (max_error < error)
@@ -258,10 +258,10 @@ class Trans4x4TestBase {

      for (int j = 0; j < kNumCoeffs; ++j) {
 #if CONFIG_VP9_HIGHBITDEPTH
-        const uint32_t diff =
+        const int diff =
            bit_depth_ == VPX_BITS_8 ? dst[j] - src[j] : dst16[j] - src16[j];
 #else
-        const uint32_t diff = dst[j] - src[j];
+        const int diff = dst[j] - src[j];
 #endif
        const uint32_t error = diff * diff;
        EXPECT_GE(static_cast<uint32_t>(limit), error)
@@ -487,19 +487,11 @@ INSTANTIATE_TEST_CASE_P(
        make_tuple(&vp9_fht4x4_c, &vp9_iht4x4_16_add_neon, 3, VPX_BITS_8)));
 #endif  // HAVE_NEON && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE

-#if CONFIG_USE_X86INC && HAVE_MMX && !CONFIG_VP9_HIGHBITDEPTH && \
-    !CONFIG_EMULATE_HARDWARE
-INSTANTIATE_TEST_CASE_P(
-    MMX, Trans4x4WHT,
-    ::testing::Values(
-        make_tuple(&vp9_fwht4x4_mmx, &vpx_iwht4x4_16_add_c, 0, VPX_BITS_8)));
-#endif
-
-#if CONFIG_USE_X86INC && HAVE_SSE2 && !CONFIG_VP9_HIGHBITDEPTH && \
-    !CONFIG_EMULATE_HARDWARE
+#if CONFIG_USE_X86INC && HAVE_SSE2 && !CONFIG_EMULATE_HARDWARE
 INSTANTIATE_TEST_CASE_P(
    SSE2, Trans4x4WHT,
    ::testing::Values(
+        make_tuple(&vp9_fwht4x4_sse2, &vpx_iwht4x4_16_add_c, 0, VPX_BITS_8),
        make_tuple(&vp9_fwht4x4_c, &vpx_iwht4x4_16_add_sse2, 0, VPX_BITS_8)));
 #endif

--- a/test/fdct8x8_test.cc
+++ b/test/fdct8x8_test.cc
@@ -47,7 +47,7 @@ typedef std::tr1::tuple<FdctFunc, IdctFunc, int, vpx_bit_depth_t> Dct8x8Param;
 typedef std::tr1::tuple<FhtFunc, IhtFunc, int, vpx_bit_depth_t> Ht8x8Param;
 typedef std::tr1::tuple<IdctFunc, IdctFunc, int, vpx_bit_depth_t> Idct8x8Param;

-void reference_8x8_dct_1d(const double in[8], double out[8], int stride) {
+void reference_8x8_dct_1d(const double in[8], double out[8]) {
  const double kInvSqrt2 = 0.707106781186547524400844362104;
  for (int k = 0; k < 8; k++) {
    out[k] = 0.0;
@@ -65,7 +65,7 @@ void reference_8x8_dct_2d(const int16_t input[kNumCoeffs],
    double temp_in[8], temp_out[8];
    for (int j = 0; j < 8; ++j)
      temp_in[j] = input[j*8 + i];
-    reference_8x8_dct_1d(temp_in, temp_out, 1);
+    reference_8x8_dct_1d(temp_in, temp_out);
    for (int j = 0; j < 8; ++j)
      output[j * 8 + i] = temp_out[j];
  }
@@ -74,7 +74,7 @@ void reference_8x8_dct_2d(const int16_t input[kNumCoeffs],
    double temp_in[8], temp_out[8];
    for (int j = 0; j < 8; ++j)
      temp_in[j] = output[j + i*8];
-    reference_8x8_dct_1d(temp_in, temp_out, 1);
+    reference_8x8_dct_1d(temp_in, temp_out);
    // Scale by some magic number
    for (int j = 0; j < 8; ++j)
      output[j + i * 8] = temp_out[j] * 2;
@@ -82,7 +82,8 @@ void reference_8x8_dct_2d(const int16_t input[kNumCoeffs],
 }


-void fdct8x8_ref(const int16_t *in, tran_low_t *out, int stride, int tx_type) {
+void fdct8x8_ref(const int16_t *in, tran_low_t *out, int stride,
+                 int /*tx_type*/) {
  vpx_fdct8x8_c(in, out, stride);
 }

@@ -107,6 +108,8 @@ void iht8x8_12(const tran_low_t *in, uint8_t *out, int stride, int tx_type) {
  vp9_highbd_iht8x8_64_add_c(in, out, stride, tx_type, 12);
 }

+#if HAVE_SSE2
+
 void idct8x8_10_add_10_c(const tran_low_t *in, uint8_t *out, int stride) {
  vpx_highbd_idct8x8_10_add_c(in, out, stride, 10);
 }
@@ -115,7 +118,6 @@ void idct8x8_10_add_12_c(const tran_low_t *in, uint8_t *out, int stride) {
  vpx_highbd_idct8x8_10_add_c(in, out, stride, 12);
 }

-#if HAVE_SSE2
 void idct8x8_10_add_10_sse2(const tran_low_t *in, uint8_t *out, int stride) {
  vpx_highbd_idct8x8_10_add_sse2(in, out, stride, 10);
 }
@@ -423,10 +425,10 @@ class FwdTrans8x8TestBase {

      for (int j = 0; j < kNumCoeffs; ++j) {
 #if CONFIG_VP9_HIGHBITDEPTH
-        const uint32_t diff =
+        const int diff =
            bit_depth_ == VPX_BITS_8 ? dst[j] - src[j] : dst16[j] - src16[j];
 #else
-        const uint32_t diff = dst[j] - src[j];
+        const int diff = dst[j] - src[j];
 #endif
        const uint32_t error = diff * diff;
        EXPECT_GE(1u << 2 * (bit_depth_ - 8), error)
@@ -456,7 +458,7 @@ class FwdTrans8x8TestBase {
        coeff_r[j] = static_cast<tran_low_t>(round(out_r[j]));

      for (int j = 0; j < kNumCoeffs; ++j) {
-        const uint32_t diff = coeff[j] - coeff_r[j];
+        const int32_t diff = coeff[j] - coeff_r[j];
        const uint32_t error = diff * diff;
        EXPECT_GE(9u << 2 * (bit_depth_ - 8), error)
            << "Error: 8x8 DCT has error " << error
@@ -509,10 +511,10 @@ void CompareInvReference(IdctFunc ref_txfm, int thresh) {

      for (int j = 0; j < kNumCoeffs; ++j) {
 #if CONFIG_VP9_HIGHBITDEPTH
-        const uint32_t diff =
+        const int diff =
            bit_depth_ == VPX_BITS_8 ? dst[j] - ref[j] : dst16[j] - ref16[j];
 #else
-        const uint32_t diff = dst[j] - ref[j];
+        const int diff = dst[j] - ref[j];
 #endif
        const uint32_t error = diff * diff;
        EXPECT_EQ(0u, error)
@@ -641,7 +643,7 @@ class InvTrans8x8DCT
  void RunInvTxfm(tran_low_t *out, uint8_t *dst, int stride) {
    inv_txfm_(out, dst, stride);
  }
-  void RunFwdTxfm(int16_t *out, tran_low_t *dst, int stride) {}
+  void RunFwdTxfm(int16_t * /*out*/, tran_low_t * /*dst*/, int /*stride*/) {}

  IdctFunc ref_txfm_;
  IdctFunc inv_txfm_;
--- a/test/hadamard_test.cc
+++ b/test/hadamard_test.cc
@@ -0,0 +1,220 @@
+/*
+ *  Copyright (c) 2016 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 <algorithm>
+
+#include "third_party/googletest/src/include/gtest/gtest.h"
+
+#include "./vpx_dsp_rtcd.h"
+
+#include "test/acm_random.h"
+#include "test/register_state_check.h"
+
+namespace {
+
+using ::libvpx_test::ACMRandom;
+
+typedef void (*HadamardFunc)(const int16_t *a, int a_stride, int16_t *b);
+
+void hadamard_loop(const int16_t *a, int a_stride, int16_t *out) {
+  int16_t b[8];
+  for (int i = 0; i < 8; i += 2) {
+    b[i + 0] = a[i * a_stride] + a[(i + 1) * a_stride];
+    b[i + 1] = a[i * a_stride] - a[(i + 1) * a_stride];
+  }
+  int16_t c[8];
+  for (int i = 0; i < 8; i += 4) {
+    c[i + 0] = b[i + 0] + b[i + 2];
+    c[i + 1] = b[i + 1] + b[i + 3];
+    c[i + 2] = b[i + 0] - b[i + 2];
+    c[i + 3] = b[i + 1] - b[i + 3];
+  }
+  out[0] = c[0] + c[4];
+  out[7] = c[1] + c[5];
+  out[3] = c[2] + c[6];
+  out[4] = c[3] + c[7];
+  out[2] = c[0] - c[4];
+  out[6] = c[1] - c[5];
+  out[1] = c[2] - c[6];
+  out[5] = c[3] - c[7];
+}
+
+void reference_hadamard8x8(const int16_t *a, int a_stride, int16_t *b) {
+  int16_t buf[64];
+  for (int i = 0; i < 8; ++i) {
+    hadamard_loop(a + i, a_stride, buf + i * 8);
+  }
+
+  for (int i = 0; i < 8; ++i) {
+    hadamard_loop(buf + i, 8, b + i * 8);
+  }
+}
+
+void reference_hadamard16x16(const int16_t *a, int a_stride, int16_t *b) {
+  /* The source is a 16x16 block. The destination is rearranged to 8x32.
+   * Input is 9 bit. */
+  reference_hadamard8x8(a + 0 + 0 * a_stride, a_stride, b + 0);
+  reference_hadamard8x8(a + 8 + 0 * a_stride, a_stride, b + 64);
+  reference_hadamard8x8(a + 0 + 8 * a_stride, a_stride, b + 128);
+  reference_hadamard8x8(a + 8 + 8 * a_stride, a_stride, b + 192);
+
+  /* Overlay the 8x8 blocks and combine. */
+  for (int i = 0; i < 64; ++i) {
+    /* 8x8 steps the range up to 15 bits. */
+    const int16_t a0 = b[0];
+    const int16_t a1 = b[64];
+    const int16_t a2 = b[128];
+    const int16_t a3 = b[192];
+
+    /* Prevent the result from escaping int16_t. */
+    const int16_t b0 = (a0 + a1) >> 1;
+    const int16_t b1 = (a0 - a1) >> 1;
+    const int16_t b2 = (a2 + a3) >> 1;
+    const int16_t b3 = (a2 - a3) >> 1;
+
+    /* Store a 16 bit value. */
+    b[  0] = b0 + b2;
+    b[ 64] = b1 + b3;
+    b[128] = b0 - b2;
+    b[192] = b1 - b3;
+
+    ++b;
+  }
+}
+
+class HadamardTestBase : public ::testing::TestWithParam<HadamardFunc> {
+ public:
+  virtual void SetUp() {
+    h_func_ = GetParam();
+    rnd_.Reset(ACMRandom::DeterministicSeed());
+  }
+
+ protected:
+  HadamardFunc h_func_;
+  ACMRandom rnd_;
+};
+
+class Hadamard8x8Test : public HadamardTestBase {};
+
+TEST_P(Hadamard8x8Test, CompareReferenceRandom) {
+  DECLARE_ALIGNED(16, int16_t, a[64]);
+  DECLARE_ALIGNED(16, int16_t, b[64]);
+  int16_t b_ref[64];
+  for (int i = 0; i < 64; ++i) {
+    a[i] = rnd_.Rand9Signed();
+  }
+  memset(b, 0, sizeof(b));
+  memset(b_ref, 0, sizeof(b_ref));
+
+  reference_hadamard8x8(a, 8, b_ref);
+  ASM_REGISTER_STATE_CHECK(h_func_(a, 8, b));
+
+  // The order of the output is not important. Sort before checking.
+  std::sort(b, b + 64);
+  std::sort(b_ref, b_ref + 64);
+  EXPECT_EQ(0, memcmp(b, b_ref, sizeof(b)));
+}
+
+TEST_P(Hadamard8x8Test, VaryStride) {
+  DECLARE_ALIGNED(16, int16_t, a[64 * 8]);
+  DECLARE_ALIGNED(16, int16_t, b[64]);
+  int16_t b_ref[64];
+  for (int i = 0; i < 64 * 8; ++i) {
+    a[i] = rnd_.Rand9Signed();
+  }
+
+  for (int i = 8; i < 64; i += 8) {
+    memset(b, 0, sizeof(b));
+    memset(b_ref, 0, sizeof(b_ref));
+
+    reference_hadamard8x8(a, i, b_ref);
+    ASM_REGISTER_STATE_CHECK(h_func_(a, i, b));
+
+    // The order of the output is not important. Sort before checking.
+    std::sort(b, b + 64);
+    std::sort(b_ref, b_ref + 64);
+    EXPECT_EQ(0, memcmp(b, b_ref, sizeof(b)));
+  }
+}
+
+INSTANTIATE_TEST_CASE_P(C, Hadamard8x8Test,
+                        ::testing::Values(&vpx_hadamard_8x8_c));
+
+#if HAVE_SSE2
+INSTANTIATE_TEST_CASE_P(SSE2, Hadamard8x8Test,
+                        ::testing::Values(&vpx_hadamard_8x8_sse2));
+#endif  // HAVE_SSE2
+
+#if HAVE_SSSE3 && CONFIG_USE_X86INC && ARCH_X86_64
+INSTANTIATE_TEST_CASE_P(SSSE3, Hadamard8x8Test,
+                        ::testing::Values(&vpx_hadamard_8x8_ssse3));
+#endif  // HAVE_SSSE3 && CONFIG_USE_X86INC && ARCH_X86_64
+
+#if HAVE_NEON
+INSTANTIATE_TEST_CASE_P(NEON, Hadamard8x8Test,
+                        ::testing::Values(&vpx_hadamard_8x8_neon));
+#endif  // HAVE_NEON
+
+class Hadamard16x16Test : public HadamardTestBase {};
+
+TEST_P(Hadamard16x16Test, CompareReferenceRandom) {
+  DECLARE_ALIGNED(16, int16_t, a[16 * 16]);
+  DECLARE_ALIGNED(16, int16_t, b[16 * 16]);
+  int16_t b_ref[16 * 16];
+  for (int i = 0; i < 16 * 16; ++i) {
+    a[i] = rnd_.Rand9Signed();
+  }
+  memset(b, 0, sizeof(b));
+  memset(b_ref, 0, sizeof(b_ref));
+
+  reference_hadamard16x16(a, 16, b_ref);
+  ASM_REGISTER_STATE_CHECK(h_func_(a, 16, b));
+
+  // The order of the output is not important. Sort before checking.
+  std::sort(b, b + 16 * 16);
+  std::sort(b_ref, b_ref + 16 * 16);
+  EXPECT_EQ(0, memcmp(b, b_ref, sizeof(b)));
+}
+
+TEST_P(Hadamard16x16Test, VaryStride) {
+  DECLARE_ALIGNED(16, int16_t, a[16 * 16 * 8]);
+  DECLARE_ALIGNED(16, int16_t, b[16 * 16]);
+  int16_t b_ref[16 * 16];
+  for (int i = 0; i < 16 * 16 * 8; ++i) {
+    a[i] = rnd_.Rand9Signed();
+  }
+
+  for (int i = 8; i < 64; i += 8) {
+    memset(b, 0, sizeof(b));
+    memset(b_ref, 0, sizeof(b_ref));
+
+    reference_hadamard16x16(a, i, b_ref);
+    ASM_REGISTER_STATE_CHECK(h_func_(a, i, b));
+
+    // The order of the output is not important. Sort before checking.
+    std::sort(b, b + 16 * 16);
+    std::sort(b_ref, b_ref + 16 * 16);
+    EXPECT_EQ(0, memcmp(b, b_ref, sizeof(b)));
+  }
+}
+
+INSTANTIATE_TEST_CASE_P(C, Hadamard16x16Test,
+                        ::testing::Values(&vpx_hadamard_16x16_c));
+
+#if HAVE_SSE2
+INSTANTIATE_TEST_CASE_P(SSE2, Hadamard16x16Test,
+                        ::testing::Values(&vpx_hadamard_16x16_sse2));
+#endif  // HAVE_SSE2
+
+#if HAVE_NEON
+INSTANTIATE_TEST_CASE_P(NEON, Hadamard16x16Test,
+                        ::testing::Values(&vpx_hadamard_16x16_neon));
+#endif  // HAVE_NEON
+}  // namespace
--- a/test/level_test.cc
+++ b/test/level_test.cc
@@ -0,0 +1,119 @@
+/*
+ *  Copyright (c) 2016 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 "third_party/googletest/src/include/gtest/gtest.h"
+#include "test/codec_factory.h"
+#include "test/encode_test_driver.h"
+#include "test/i420_video_source.h"
+#include "test/util.h"
+
+namespace {
+class LevelTest
+    : public ::libvpx_test::EncoderTest,
+      public ::libvpx_test::CodecTestWith2Params<libvpx_test::TestMode, int> {
+ protected:
+  LevelTest()
+     : EncoderTest(GET_PARAM(0)),
+       encoding_mode_(GET_PARAM(1)),
+       cpu_used_(GET_PARAM(2)),
+       min_gf_internal_(24),
+       target_level_(0),
+       level_(0) {}
+  virtual ~LevelTest() {}
+
+  virtual void SetUp() {
+    InitializeConfig();
+    SetMode(encoding_mode_);
+    if (encoding_mode_ != ::libvpx_test::kRealTime) {
+      cfg_.g_lag_in_frames = 25;
+      cfg_.rc_end_usage = VPX_VBR;
+    } else {
+      cfg_.g_lag_in_frames = 0;
+      cfg_.rc_end_usage = VPX_CBR;
+    }
+    cfg_.rc_2pass_vbr_minsection_pct = 5;
+    cfg_.rc_2pass_vbr_maxsection_pct = 2000;
+    cfg_.rc_target_bitrate = 400;
+    cfg_.rc_max_quantizer = 63;
+    cfg_.rc_min_quantizer = 0;
+  }
+
+  virtual void PreEncodeFrameHook(::libvpx_test::VideoSource *video,
+                                  ::libvpx_test::Encoder *encoder) {
+    if (video->frame() == 0) {
+      encoder->Control(VP8E_SET_CPUUSED, cpu_used_);
+      encoder->Control(VP9E_SET_TARGET_LEVEL, target_level_);
+      encoder->Control(VP9E_SET_MIN_GF_INTERVAL, min_gf_internal_);
+      if (encoding_mode_ != ::libvpx_test::kRealTime) {
+        encoder->Control(VP8E_SET_ENABLEAUTOALTREF, 1);
+        encoder->Control(VP8E_SET_ARNR_MAXFRAMES, 7);
+        encoder->Control(VP8E_SET_ARNR_STRENGTH, 5);
+        encoder->Control(VP8E_SET_ARNR_TYPE, 3);
+      }
+    }
+    encoder->Control(VP9E_GET_LEVEL, &level_);
+    ASSERT_LE(level_, 51);
+    ASSERT_GE(level_, 0);
+  }
+
+  ::libvpx_test::TestMode encoding_mode_;
+  int cpu_used_;
+  int min_gf_internal_;
+  int target_level_;
+  int level_;
+};
+
+// Test for keeping level stats only
+TEST_P(LevelTest, TestTargetLevel0) {
+  ::libvpx_test::I420VideoSource video("hantro_odd.yuv", 208, 144, 30, 1, 0,
+                                       40);
+  target_level_ = 0;
+  min_gf_internal_ = 4;
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  ASSERT_EQ(11, level_);
+
+  cfg_.rc_target_bitrate = 1600;
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  ASSERT_EQ(20, level_);
+}
+
+// Test for level control being turned off
+TEST_P(LevelTest, TestTargetLevel255) {
+  ::libvpx_test::I420VideoSource video("hantro_odd.yuv", 208, 144, 30, 1, 0,
+                                       30);
+  target_level_ = 255;
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+}
+
+TEST_P(LevelTest, TestTargetLevelApi) {
+  ::libvpx_test::I420VideoSource video("hantro_odd.yuv", 208, 144, 30, 1, 0, 1);
+  static const vpx_codec_iface_t *codec = &vpx_codec_vp9_cx_algo;
+  vpx_codec_ctx_t enc;
+  vpx_codec_enc_cfg_t cfg;
+  EXPECT_EQ(VPX_CODEC_OK, vpx_codec_enc_config_default(codec, &cfg, 0));
+  EXPECT_EQ(VPX_CODEC_OK, vpx_codec_enc_init(&enc, codec, &cfg, 0));
+  for (int level = 0; level <= 256; ++level) {
+    if (level == 10 || level == 11 || level == 20 || level == 21 ||
+        level == 30 || level == 31 || level == 40 || level == 41 ||
+        level == 50 || level == 51 || level == 52 || level == 60 ||
+        level == 61 || level == 62 || level == 0 || level == 255)
+      EXPECT_EQ(VPX_CODEC_OK,
+                vpx_codec_control(&enc, VP9E_SET_TARGET_LEVEL, level));
+    else
+      EXPECT_EQ(VPX_CODEC_INVALID_PARAM,
+                vpx_codec_control(&enc, VP9E_SET_TARGET_LEVEL, level));
+  }
+  EXPECT_EQ(VPX_CODEC_OK, vpx_codec_destroy(&enc));
+}
+
+VP9_INSTANTIATE_TEST_CASE(LevelTest,
+                          ::testing::Values(::libvpx_test::kTwoPassGood,
+                                            ::libvpx_test::kOnePassGood),
+                          ::testing::Range(0, 9));
+}  // namespace
--- a/test/lpf_8_test.cc
+++ b/test/lpf_8_test.cc
@@ -37,120 +37,23 @@ const int number_of_iterations = 10000;
 #if CONFIG_VP9_HIGHBITDEPTH
 typedef void (*loop_op_t)(uint16_t *s, int p, const uint8_t *blimit,
                          const uint8_t *limit, const uint8_t *thresh,
-                          int count, int bd);
+                          int bd);
 typedef void (*dual_loop_op_t)(uint16_t *s, int p, const uint8_t *blimit0,
                               const uint8_t *limit0, const uint8_t *thresh0,
                               const uint8_t *blimit1, const uint8_t *limit1,
                               const uint8_t *thresh1, int bd);
 #else
 typedef void (*loop_op_t)(uint8_t *s, int p, const uint8_t *blimit,
-                          const uint8_t *limit, const uint8_t *thresh,
-                          int count);
+                          const uint8_t *limit, const uint8_t *thresh);
 typedef void (*dual_loop_op_t)(uint8_t *s, int p, const uint8_t *blimit0,
                               const uint8_t *limit0, const uint8_t *thresh0,
                               const uint8_t *blimit1, const uint8_t *limit1,
                               const uint8_t *thresh1);
 #endif  // CONFIG_VP9_HIGHBITDEPTH

-typedef std::tr1::tuple<loop_op_t, loop_op_t, int, int> loop8_param_t;
+typedef std::tr1::tuple<loop_op_t, loop_op_t, int> loop8_param_t;
 typedef std::tr1::tuple<dual_loop_op_t, dual_loop_op_t, int> dualloop8_param_t;

-#if HAVE_SSE2
-#if CONFIG_VP9_HIGHBITDEPTH
-void wrapper_vertical_16_sse2(uint16_t *s, int p, const uint8_t *blimit,
-                              const uint8_t *limit, const uint8_t *thresh,
-                              int count, int bd) {
-  vpx_highbd_lpf_vertical_16_sse2(s, p, blimit, limit, thresh, bd);
-}
-
-void wrapper_vertical_16_c(uint16_t *s, int p, const uint8_t *blimit,
-                           const uint8_t *limit, const uint8_t *thresh,
-                           int count, int bd) {
-  vpx_highbd_lpf_vertical_16_c(s, p, blimit, limit, thresh, bd);
-}
-
-void wrapper_vertical_16_dual_sse2(uint16_t *s, int p, const uint8_t *blimit,
-                                   const uint8_t *limit, const uint8_t *thresh,
-                                   int count, int bd) {
-  vpx_highbd_lpf_vertical_16_dual_sse2(s, p, blimit, limit, thresh, bd);
-}
-
-void wrapper_vertical_16_dual_c(uint16_t *s, int p, const uint8_t *blimit,
-                                const uint8_t *limit, const uint8_t *thresh,
-                                int count, int bd) {
-  vpx_highbd_lpf_vertical_16_dual_c(s, p, blimit, limit, thresh, bd);
-}
-#else
-void wrapper_vertical_16_sse2(uint8_t *s, int p, const uint8_t *blimit,
-                              const uint8_t *limit, const uint8_t *thresh,
-                              int count) {
-  vpx_lpf_vertical_16_sse2(s, p, blimit, limit, thresh);
-}
-
-void wrapper_vertical_16_c(uint8_t *s, int p, const uint8_t *blimit,
-                           const uint8_t *limit, const uint8_t *thresh,
-                           int count) {
-  vpx_lpf_vertical_16_c(s, p, blimit, limit, thresh);
-}
-
-void wrapper_vertical_16_dual_sse2(uint8_t *s, int p, const uint8_t *blimit,
-                                   const uint8_t *limit, const uint8_t *thresh,
-                                   int count) {
-  vpx_lpf_vertical_16_dual_sse2(s, p, blimit, limit, thresh);
-}
-
-void wrapper_vertical_16_dual_c(uint8_t *s, int p, const uint8_t *blimit,
-                                const uint8_t *limit, const uint8_t *thresh,
-                                int count) {
-  vpx_lpf_vertical_16_dual_c(s, p, blimit, limit, thresh);
-}
-#endif  // CONFIG_VP9_HIGHBITDEPTH
-#endif  // HAVE_SSE2
-
-#if HAVE_NEON_ASM
-#if CONFIG_VP9_HIGHBITDEPTH
-// No neon high bitdepth functions.
-#else
-void wrapper_vertical_16_neon(uint8_t *s, int p, const uint8_t *blimit,
-                              const uint8_t *limit, const uint8_t *thresh,
-                              int count) {
-  vpx_lpf_vertical_16_neon(s, p, blimit, limit, thresh);
-}
-
-void wrapper_vertical_16_c(uint8_t *s, int p, const uint8_t *blimit,
-                           const uint8_t *limit, const uint8_t *thresh,
-                           int count) {
-  vpx_lpf_vertical_16_c(s, p, blimit, limit, thresh);
-}
-
-void wrapper_vertical_16_dual_neon(uint8_t *s, int p, const uint8_t *blimit,
-                                   const uint8_t *limit, const uint8_t *thresh,
-                                   int count) {
-  vpx_lpf_vertical_16_dual_neon(s, p, blimit, limit, thresh);
-}
-
-void wrapper_vertical_16_dual_c(uint8_t *s, int p, const uint8_t *blimit,
-                                const uint8_t *limit, const uint8_t *thresh,
-                                int count) {
-  vpx_lpf_vertical_16_dual_c(s, p, blimit, limit, thresh);
-}
-#endif  // CONFIG_VP9_HIGHBITDEPTH
-#endif  // HAVE_NEON_ASM
-
-#if HAVE_MSA && (!CONFIG_VP9_HIGHBITDEPTH)
-void wrapper_vertical_16_msa(uint8_t *s, int p, const uint8_t *blimit,
-                             const uint8_t *limit, const uint8_t *thresh,
-                             int count) {
-  vpx_lpf_vertical_16_msa(s, p, blimit, limit, thresh);
-}
-
-void wrapper_vertical_16_c(uint8_t *s, int p, const uint8_t *blimit,
-                           const uint8_t *limit, const uint8_t *thresh,
-                           int count) {
-  vpx_lpf_vertical_16_c(s, p, blimit, limit, thresh);
-}
-#endif  // HAVE_MSA && (!CONFIG_VP9_HIGHBITDEPTH)
-
 class Loop8Test6Param : public ::testing::TestWithParam<loop8_param_t> {
 public:
  virtual ~Loop8Test6Param() {}
@@ -158,7 +61,6 @@ class Loop8Test6Param : public ::testing::TestWithParam<loop8_param_t> {
    loopfilter_op_ = GET_PARAM(0);
    ref_loopfilter_op_ = GET_PARAM(1);
    bit_depth_ = GET_PARAM(2);
-    count_ = GET_PARAM(3);
    mask_ = (1 << bit_depth_) - 1;
  }

@@ -166,7 +68,6 @@ class Loop8Test6Param : public ::testing::TestWithParam<loop8_param_t> {

 protected:
  int bit_depth_;
-  int count_;
  int mask_;
  loop_op_t loopfilter_op_;
  loop_op_t ref_loopfilter_op_;
@@ -253,13 +154,13 @@ TEST_P(Loop8Test6Param, OperationCheck) {
      ref_s[j] = s[j];
    }
 #if CONFIG_VP9_HIGHBITDEPTH
-    ref_loopfilter_op_(ref_s + 8 + p * 8, p, blimit, limit, thresh, count_, bd);
+    ref_loopfilter_op_(ref_s + 8 + p * 8, p, blimit, limit, thresh, bd);
    ASM_REGISTER_STATE_CHECK(
-        loopfilter_op_(s + 8 + p * 8, p, blimit, limit, thresh, count_, bd));
+        loopfilter_op_(s + 8 + p * 8, p, blimit, limit, thresh, bd));
 #else
-    ref_loopfilter_op_(ref_s+8+p*8, p, blimit, limit, thresh, count_);
+    ref_loopfilter_op_(ref_s+8+p*8, p, blimit, limit, thresh);
    ASM_REGISTER_STATE_CHECK(
-        loopfilter_op_(s + 8 + p * 8, p, blimit, limit, thresh, count_));
+        loopfilter_op_(s + 8 + p * 8, p, blimit, limit, thresh));
 #endif  // CONFIG_VP9_HIGHBITDEPTH

    for (int j = 0; j < kNumCoeffs; ++j) {
@@ -325,13 +226,13 @@ TEST_P(Loop8Test6Param, ValueCheck) {
      ref_s[j] = s[j];
    }
 #if CONFIG_VP9_HIGHBITDEPTH
-    ref_loopfilter_op_(ref_s + 8 + p * 8, p, blimit, limit, thresh, count_, bd);
+    ref_loopfilter_op_(ref_s + 8 + p * 8, p, blimit, limit, thresh, bd);
    ASM_REGISTER_STATE_CHECK(
-        loopfilter_op_(s + 8 + p * 8, p, blimit, limit, thresh, count_, bd));
+        loopfilter_op_(s + 8 + p * 8, p, blimit, limit, thresh, bd));
 #else
-    ref_loopfilter_op_(ref_s+8+p*8, p, blimit, limit, thresh, count_);
+    ref_loopfilter_op_(ref_s+8+p*8, p, blimit, limit, thresh);
    ASM_REGISTER_STATE_CHECK(
-        loopfilter_op_(s + 8 + p * 8, p, blimit, limit, thresh, count_));
+        loopfilter_op_(s + 8 + p * 8, p, blimit, limit, thresh));
 #endif  // CONFIG_VP9_HIGHBITDEPTH
    for (int j = 0; j < kNumCoeffs; ++j) {
      err_count += ref_s[j] != s[j];
@@ -535,64 +436,73 @@ INSTANTIATE_TEST_CASE_P(
    SSE2, Loop8Test6Param,
    ::testing::Values(
        make_tuple(&vpx_highbd_lpf_horizontal_4_sse2,
-                   &vpx_highbd_lpf_horizontal_4_c, 8, 1),
+                   &vpx_highbd_lpf_horizontal_4_c, 8),
        make_tuple(&vpx_highbd_lpf_vertical_4_sse2,
-                   &vpx_highbd_lpf_vertical_4_c, 8, 1),
+                   &vpx_highbd_lpf_vertical_4_c, 8),
        make_tuple(&vpx_highbd_lpf_horizontal_8_sse2,
-                   &vpx_highbd_lpf_horizontal_8_c, 8, 1),
-        make_tuple(&vpx_highbd_lpf_horizontal_16_sse2,
-                   &vpx_highbd_lpf_horizontal_16_c, 8, 1),
-        make_tuple(&vpx_highbd_lpf_horizontal_16_sse2,
-                   &vpx_highbd_lpf_horizontal_16_c, 8, 2),
+                   &vpx_highbd_lpf_horizontal_8_c, 8),
+        make_tuple(&vpx_highbd_lpf_horizontal_edge_8_sse2,
+                   &vpx_highbd_lpf_horizontal_edge_8_c, 8),
+        make_tuple(&vpx_highbd_lpf_horizontal_edge_16_sse2,
+                   &vpx_highbd_lpf_horizontal_edge_16_c, 8),
        make_tuple(&vpx_highbd_lpf_vertical_8_sse2,
-                   &vpx_highbd_lpf_vertical_8_c, 8, 1),
-        make_tuple(&wrapper_vertical_16_sse2,
-                   &wrapper_vertical_16_c, 8, 1),
+                   &vpx_highbd_lpf_vertical_8_c, 8),
+        make_tuple(&vpx_highbd_lpf_vertical_16_sse2,
+                   &vpx_highbd_lpf_vertical_16_c, 8),
        make_tuple(&vpx_highbd_lpf_horizontal_4_sse2,
-                   &vpx_highbd_lpf_horizontal_4_c, 10, 1),
+                   &vpx_highbd_lpf_horizontal_4_c, 10),
        make_tuple(&vpx_highbd_lpf_vertical_4_sse2,
-                   &vpx_highbd_lpf_vertical_4_c, 10, 1),
+                   &vpx_highbd_lpf_vertical_4_c, 10),
        make_tuple(&vpx_highbd_lpf_horizontal_8_sse2,
-                   &vpx_highbd_lpf_horizontal_8_c, 10, 1),
-        make_tuple(&vpx_highbd_lpf_horizontal_16_sse2,
-                   &vpx_highbd_lpf_horizontal_16_c, 10, 1),
-        make_tuple(&vpx_highbd_lpf_horizontal_16_sse2,
-                   &vpx_highbd_lpf_horizontal_16_c, 10, 2),
+                   &vpx_highbd_lpf_horizontal_8_c, 10),
+        make_tuple(&vpx_highbd_lpf_horizontal_edge_8_sse2,
+                   &vpx_highbd_lpf_horizontal_edge_8_c, 10),
+        make_tuple(&vpx_highbd_lpf_horizontal_edge_16_sse2,
+                   &vpx_highbd_lpf_horizontal_edge_16_c, 10),
        make_tuple(&vpx_highbd_lpf_vertical_8_sse2,
-                   &vpx_highbd_lpf_vertical_8_c, 10, 1),
-        make_tuple(&wrapper_vertical_16_sse2,
-                   &wrapper_vertical_16_c, 10, 1),
+                   &vpx_highbd_lpf_vertical_8_c, 10),
+        make_tuple(&vpx_highbd_lpf_vertical_16_sse2,
+                   &vpx_highbd_lpf_vertical_16_c, 10),
        make_tuple(&vpx_highbd_lpf_horizontal_4_sse2,
-                   &vpx_highbd_lpf_horizontal_4_c, 12, 1),
+                   &vpx_highbd_lpf_horizontal_4_c, 12),
        make_tuple(&vpx_highbd_lpf_vertical_4_sse2,
-                   &vpx_highbd_lpf_vertical_4_c, 12, 1),
+                   &vpx_highbd_lpf_vertical_4_c, 12),
        make_tuple(&vpx_highbd_lpf_horizontal_8_sse2,
-                   &vpx_highbd_lpf_horizontal_8_c, 12, 1),
-        make_tuple(&vpx_highbd_lpf_horizontal_16_sse2,
-                   &vpx_highbd_lpf_horizontal_16_c, 12, 1),
-        make_tuple(&vpx_highbd_lpf_horizontal_16_sse2,
-                   &vpx_highbd_lpf_horizontal_16_c, 12, 2),
+                   &vpx_highbd_lpf_horizontal_8_c, 12),
+        make_tuple(&vpx_highbd_lpf_horizontal_edge_8_sse2,
+                   &vpx_highbd_lpf_horizontal_edge_8_c, 12),
+        make_tuple(&vpx_highbd_lpf_horizontal_edge_16_sse2,
+                   &vpx_highbd_lpf_horizontal_edge_16_c, 12),
        make_tuple(&vpx_highbd_lpf_vertical_8_sse2,
-                   &vpx_highbd_lpf_vertical_8_c, 12, 1),
-        make_tuple(&wrapper_vertical_16_sse2,
-                   &wrapper_vertical_16_c, 12, 1),
-        make_tuple(&wrapper_vertical_16_dual_sse2,
-                   &wrapper_vertical_16_dual_c, 8, 1),
-        make_tuple(&wrapper_vertical_16_dual_sse2,
-                   &wrapper_vertical_16_dual_c, 10, 1),
-        make_tuple(&wrapper_vertical_16_dual_sse2,
-                   &wrapper_vertical_16_dual_c, 12, 1)));
+                   &vpx_highbd_lpf_vertical_8_c, 12),
+        make_tuple(&vpx_highbd_lpf_vertical_16_sse2,
+                   &vpx_highbd_lpf_vertical_16_c, 12),
+        make_tuple(&vpx_highbd_lpf_vertical_16_dual_sse2,
+                   &vpx_highbd_lpf_vertical_16_dual_c, 8),
+        make_tuple(&vpx_highbd_lpf_vertical_16_dual_sse2,
+                   &vpx_highbd_lpf_vertical_16_dual_c, 10),
+        make_tuple(&vpx_highbd_lpf_vertical_16_dual_sse2,
+                   &vpx_highbd_lpf_vertical_16_dual_c, 12)));
 #else
 INSTANTIATE_TEST_CASE_P(
    SSE2, Loop8Test6Param,
    ::testing::Values(
-        make_tuple(&vpx_lpf_horizontal_8_sse2, &vpx_lpf_horizontal_8_c, 8, 1),
-        make_tuple(&vpx_lpf_horizontal_16_sse2, &vpx_lpf_horizontal_16_c, 8, 1),
-        make_tuple(&vpx_lpf_horizontal_16_sse2, &vpx_lpf_horizontal_16_c, 8, 2),
-        make_tuple(&vpx_lpf_vertical_8_sse2, &vpx_lpf_vertical_8_c, 8, 1),
-        make_tuple(&wrapper_vertical_16_sse2, &wrapper_vertical_16_c, 8, 1),
-        make_tuple(&wrapper_vertical_16_dual_sse2,
-                   &wrapper_vertical_16_dual_c, 8, 1)));
+        make_tuple(&vpx_lpf_horizontal_4_sse2,
+                   &vpx_lpf_horizontal_4_c, 8),
+        make_tuple(&vpx_lpf_horizontal_8_sse2,
+                   &vpx_lpf_horizontal_8_c, 8),
+        make_tuple(&vpx_lpf_horizontal_edge_8_sse2,
+                   &vpx_lpf_horizontal_edge_8_c, 8),
+        make_tuple(&vpx_lpf_horizontal_edge_16_sse2,
+                   &vpx_lpf_horizontal_edge_16_c, 8),
+        make_tuple(&vpx_lpf_vertical_4_sse2,
+                   &vpx_lpf_vertical_4_c, 8),
+        make_tuple(&vpx_lpf_vertical_8_sse2,
+                   &vpx_lpf_vertical_8_c, 8),
+        make_tuple(&vpx_lpf_vertical_16_sse2,
+                   &vpx_lpf_vertical_16_c, 8),
+        make_tuple(&vpx_lpf_vertical_16_dual_sse2,
+                   &vpx_lpf_vertical_16_dual_c, 8)));
 #endif  // CONFIG_VP9_HIGHBITDEPTH
 #endif

@@ -600,9 +510,10 @@ INSTANTIATE_TEST_CASE_P(
 INSTANTIATE_TEST_CASE_P(
    AVX2, Loop8Test6Param,
    ::testing::Values(
-        make_tuple(&vpx_lpf_horizontal_16_avx2, &vpx_lpf_horizontal_16_c, 8, 1),
-        make_tuple(&vpx_lpf_horizontal_16_avx2, &vpx_lpf_horizontal_16_c, 8,
-                   2)));
+        make_tuple(&vpx_lpf_horizontal_edge_8_avx2,
+                   &vpx_lpf_horizontal_edge_8_c, 8),
+        make_tuple(&vpx_lpf_horizontal_edge_16_avx2,
+                   &vpx_lpf_horizontal_edge_16_c, 8)));
 #endif

 #if HAVE_SSE2
@@ -659,23 +570,23 @@ INSTANTIATE_TEST_CASE_P(
 #if HAVE_NEON_ASM
 // Using #if inside the macro is unsupported on MSVS but the tests are not
 // currently built for MSVS with ARM and NEON.
-        make_tuple(&vpx_lpf_horizontal_16_neon,
-                   &vpx_lpf_horizontal_16_c, 8, 1),
-        make_tuple(&vpx_lpf_horizontal_16_neon,
-                   &vpx_lpf_horizontal_16_c, 8, 2),
-        make_tuple(&wrapper_vertical_16_neon,
-                   &wrapper_vertical_16_c, 8, 1),
-        make_tuple(&wrapper_vertical_16_dual_neon,
-                   &wrapper_vertical_16_dual_c, 8, 1),
+        make_tuple(&vpx_lpf_horizontal_edge_8_neon,
+                   &vpx_lpf_horizontal_edge_8_c, 8),
+        make_tuple(&vpx_lpf_horizontal_edge_16_neon,
+                   &vpx_lpf_horizontal_edge_16_c, 8),
+        make_tuple(&vpx_lpf_vertical_16_neon,
+                   &vpx_lpf_vertical_16_c, 8),
+        make_tuple(&vpx_lpf_vertical_16_dual_neon,
+                   &vpx_lpf_vertical_16_dual_c, 8),
 #endif  // HAVE_NEON_ASM
        make_tuple(&vpx_lpf_horizontal_8_neon,
-                   &vpx_lpf_horizontal_8_c, 8, 1),
+                   &vpx_lpf_horizontal_8_c, 8),
        make_tuple(&vpx_lpf_vertical_8_neon,
-                   &vpx_lpf_vertical_8_c, 8, 1),
+                   &vpx_lpf_vertical_8_c, 8),
        make_tuple(&vpx_lpf_horizontal_4_neon,
-                   &vpx_lpf_horizontal_4_c, 8, 1),
+                   &vpx_lpf_horizontal_4_c, 8),
        make_tuple(&vpx_lpf_vertical_4_neon,
-                   &vpx_lpf_vertical_4_c, 8, 1)));
+                   &vpx_lpf_vertical_4_c, 8)));
 INSTANTIATE_TEST_CASE_P(
    NEON, Loop8Test9Param,
    ::testing::Values(
@@ -692,15 +603,58 @@ INSTANTIATE_TEST_CASE_P(
 #endif  // CONFIG_VP9_HIGHBITDEPTH
 #endif  // HAVE_NEON

+#if HAVE_DSPR2 && !CONFIG_VP9_HIGHBITDEPTH
+INSTANTIATE_TEST_CASE_P(
+    DSPR2, Loop8Test6Param,
+    ::testing::Values(
+        make_tuple(&vpx_lpf_horizontal_4_dspr2,
+                   &vpx_lpf_horizontal_4_c, 8),
+        make_tuple(&vpx_lpf_horizontal_8_dspr2,
+                   &vpx_lpf_horizontal_8_c, 8),
+        make_tuple(&vpx_lpf_horizontal_edge_8,
+                   &vpx_lpf_horizontal_edge_8, 8),
+        make_tuple(&vpx_lpf_horizontal_edge_16,
+                   &vpx_lpf_horizontal_edge_16, 8),
+        make_tuple(&vpx_lpf_vertical_4_dspr2,
+                   &vpx_lpf_vertical_4_c, 8),
+        make_tuple(&vpx_lpf_vertical_8_dspr2,
+                   &vpx_lpf_vertical_8_c, 8),
+        make_tuple(&vpx_lpf_vertical_16_dspr2,
+                   &vpx_lpf_vertical_16_c, 8),
+        make_tuple(&vpx_lpf_vertical_16_dual_dspr2,
+                   &vpx_lpf_vertical_16_dual_c, 8)));
+
+INSTANTIATE_TEST_CASE_P(
+    DSPR2, Loop8Test9Param,
+    ::testing::Values(
+        make_tuple(&vpx_lpf_horizontal_4_dual_dspr2,
+                   &vpx_lpf_horizontal_4_dual_c, 8),
+        make_tuple(&vpx_lpf_horizontal_8_dual_dspr2,
+                   &vpx_lpf_horizontal_8_dual_c, 8),
+        make_tuple(&vpx_lpf_vertical_4_dual_dspr2,
+                   &vpx_lpf_vertical_4_dual_c, 8),
+        make_tuple(&vpx_lpf_vertical_8_dual_dspr2,
+                   &vpx_lpf_vertical_8_dual_c, 8)));
+#endif  // HAVE_DSPR2 && !CONFIG_VP9_HIGHBITDEPTH
+
 #if HAVE_MSA && (!CONFIG_VP9_HIGHBITDEPTH)
 INSTANTIATE_TEST_CASE_P(
    MSA, Loop8Test6Param,
    ::testing::Values(
-        make_tuple(&vpx_lpf_horizontal_8_msa, &vpx_lpf_horizontal_8_c, 8, 1),
-        make_tuple(&vpx_lpf_horizontal_16_msa, &vpx_lpf_horizontal_16_c, 8, 1),
-        make_tuple(&vpx_lpf_horizontal_16_msa, &vpx_lpf_horizontal_16_c, 8, 2),
-        make_tuple(&vpx_lpf_vertical_8_msa, &vpx_lpf_vertical_8_c, 8, 1),
-        make_tuple(&wrapper_vertical_16_msa, &wrapper_vertical_16_c, 8, 1)));
+        make_tuple(&vpx_lpf_horizontal_4_msa,
+                   &vpx_lpf_horizontal_4_c, 8),
+        make_tuple(&vpx_lpf_horizontal_8_msa,
+                   &vpx_lpf_horizontal_8_c, 8),
+        make_tuple(&vpx_lpf_horizontal_edge_8_msa,
+                   &vpx_lpf_horizontal_edge_8_c, 8),
+        make_tuple(&vpx_lpf_horizontal_edge_16_msa,
+                   &vpx_lpf_horizontal_edge_16_c, 8),
+        make_tuple(&vpx_lpf_vertical_4_msa,
+                   &vpx_lpf_vertical_4_c, 8),
+        make_tuple(&vpx_lpf_vertical_8_msa,
+                   &vpx_lpf_vertical_8_c, 8),
+        make_tuple(&vpx_lpf_vertical_16_msa,
+                   &vpx_lpf_vertical_16_c, 8)));

 INSTANTIATE_TEST_CASE_P(
    MSA, Loop8Test9Param,
--- a/test/minmax_test.cc
+++ b/test/minmax_test.cc
@@ -0,0 +1,132 @@
+/*
+ *  Copyright (c) 2016 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 <string.h>
+
+#include "third_party/googletest/src/include/gtest/gtest.h"
+
+#include "./vpx_dsp_rtcd.h"
+#include "vpx/vpx_integer.h"
+
+#include "test/acm_random.h"
+#include "test/register_state_check.h"
+
+namespace {
+
+using ::libvpx_test::ACMRandom;
+
+typedef void (*MinMaxFunc)(const uint8_t *a, int a_stride,
+                           const uint8_t *b, int b_stride,
+                           int *min, int *max);
+
+class MinMaxTest : public ::testing::TestWithParam<MinMaxFunc> {
+ public:
+  virtual void SetUp() {
+    mm_func_ = GetParam();
+    rnd_.Reset(ACMRandom::DeterministicSeed());
+  }
+
+ protected:
+  MinMaxFunc mm_func_;
+  ACMRandom rnd_;
+};
+
+void reference_minmax(const uint8_t *a, int a_stride,
+                      const uint8_t *b, int b_stride,
+                      int *min_ret, int *max_ret) {
+  int min = 255;
+  int max = 0;
+  for (int i = 0; i < 8; i++) {
+    for (int j = 0; j < 8; j++) {
+      const int diff = abs(a[i * a_stride + j] - b[i * b_stride + j]);
+      if (min > diff) min = diff;
+      if (max < diff) max = diff;
+    }
+  }
+
+  *min_ret = min;
+  *max_ret = max;
+}
+
+TEST_P(MinMaxTest, MinValue) {
+  for (int i = 0; i < 64; i++) {
+    uint8_t a[64], b[64];
+    memset(a, 0, sizeof(a));
+    memset(b, 255, sizeof(b));
+    b[i] = i;  // Set a minimum difference of i.
+
+    int min, max;
+    ASM_REGISTER_STATE_CHECK(mm_func_(a, 8, b, 8, &min, &max));
+    EXPECT_EQ(255, max);
+    EXPECT_EQ(i, min);
+  }
+}
+
+TEST_P(MinMaxTest, MaxValue) {
+  for (int i = 0; i < 64; i++) {
+    uint8_t a[64], b[64];
+    memset(a, 0, sizeof(a));
+    memset(b, 0, sizeof(b));
+    b[i] = i;  // Set a maximum difference of i.
+
+    int min, max;
+    ASM_REGISTER_STATE_CHECK(mm_func_(a, 8, b, 8, &min, &max));
+    EXPECT_EQ(i, max);
+    EXPECT_EQ(0, min);
+  }
+}
+
+TEST_P(MinMaxTest, CompareReference) {
+  uint8_t a[64], b[64];
+  for (int j = 0; j < 64; j++) {
+    a[j] = rnd_.Rand8();
+    b[j] = rnd_.Rand8();
+  }
+
+  int min_ref, max_ref, min, max;
+  reference_minmax(a, 8, b, 8, &min_ref, &max_ref);
+  ASM_REGISTER_STATE_CHECK(mm_func_(a, 8, b, 8, &min, &max));
+  EXPECT_EQ(max_ref, max);
+  EXPECT_EQ(min_ref, min);
+}
+
+TEST_P(MinMaxTest, CompareReferenceAndVaryStride) {
+  uint8_t a[8 * 64], b[8 * 64];
+  for (int i = 0; i < 8 * 64; i++) {
+    a[i] = rnd_.Rand8();
+    b[i] = rnd_.Rand8();
+  }
+  for (int a_stride = 8; a_stride <= 64; a_stride += 8) {
+    for (int b_stride = 8; b_stride <= 64; b_stride += 8) {
+      int min_ref, max_ref, min, max;
+      reference_minmax(a, a_stride, b, b_stride, &min_ref, &max_ref);
+      ASM_REGISTER_STATE_CHECK(mm_func_(a, a_stride, b, b_stride, &min, &max));
+      EXPECT_EQ(max_ref, max) << "when a_stride = " << a_stride
+                              << " and b_stride = " << b_stride;;
+      EXPECT_EQ(min_ref, min) << "when a_stride = " << a_stride
+                              << " and b_stride = " << b_stride;;
+    }
+  }
+}
+
+INSTANTIATE_TEST_CASE_P(C, MinMaxTest, ::testing::Values(&vpx_minmax_8x8_c));
+
+#if HAVE_SSE2
+INSTANTIATE_TEST_CASE_P(SSE2, MinMaxTest,
+                        ::testing::Values(&vpx_minmax_8x8_sse2));
+#endif
+
+#if HAVE_NEON
+INSTANTIATE_TEST_CASE_P(NEON, MinMaxTest,
+                        ::testing::Values(&vpx_minmax_8x8_neon));
+#endif
+
+}  // namespace
--- a/test/realtime_test.cc
+++ b/test/realtime_test.cc
@@ -0,0 +1,64 @@
+/*
+ *  Copyright (c) 2016 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 "test/codec_factory.h"
+#include "test/encode_test_driver.h"
+#include "test/util.h"
+#include "test/video_source.h"
+#include "third_party/googletest/src/include/gtest/gtest.h"
+
+namespace {
+
+const int kVideoSourceWidth = 320;
+const int kVideoSourceHeight = 240;
+const int kFramesToEncode = 2;
+
+class RealtimeTest
+    : public ::libvpx_test::EncoderTest,
+      public ::libvpx_test::CodecTestWithParam<libvpx_test::TestMode> {
+ protected:
+  RealtimeTest()
+      : EncoderTest(GET_PARAM(0)), frame_packets_(0) {}
+  virtual ~RealtimeTest() {}
+
+  virtual void SetUp() {
+    InitializeConfig();
+    cfg_.g_lag_in_frames = 0;
+    SetMode(::libvpx_test::kRealTime);
+  }
+
+  virtual void BeginPassHook(unsigned int /*pass*/) {
+    // TODO(tomfinegan): We're changing the pass value here to make sure
+    // we get frames when real time mode is combined with |g_pass| set to
+    // VPX_RC_FIRST_PASS. This is necessary because EncoderTest::RunLoop() sets
+    // the pass value based on the mode passed into EncoderTest::SetMode(),
+    // which overrides the one specified in SetUp() above.
+    cfg_.g_pass = VPX_RC_FIRST_PASS;
+  }
+  virtual void FramePktHook(const vpx_codec_cx_pkt_t * /*pkt*/) {
+    frame_packets_++;
+  }
+
+  int frame_packets_;
+};
+
+TEST_P(RealtimeTest, RealtimeFirstPassProducesFrames) {
+  ::libvpx_test::RandomVideoSource video;
+  video.SetSize(kVideoSourceWidth, kVideoSourceHeight);
+  video.set_limit(kFramesToEncode);
+  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
+  EXPECT_EQ(kFramesToEncode, frame_packets_);
+}
+
+VP8_INSTANTIATE_TEST_CASE(RealtimeTest,
+                          ::testing::Values(::libvpx_test::kRealTime));
+VP9_INSTANTIATE_TEST_CASE(RealtimeTest,
+                          ::testing::Values(::libvpx_test::kRealTime));
+
+}  // namespace
--- a/test/register_state_check.h
+++ b/test/register_state_check.h
@@ -36,16 +36,10 @@
 #include <windows.h>
 #include <winnt.h>

-namespace testing {
-namespace internal {
-
 inline bool operator==(const M128A& lhs, const M128A& rhs) {
  return (lhs.Low == rhs.Low && lhs.High == rhs.High);
 }

-}  // namespace internal
-}  // namespace testing
-
 namespace libvpx_test {

 // Compares the state of xmm[6-15] at construction with their state at
--- a/test/resize_test.cc
+++ b/test/resize_test.cc
@@ -7,6 +7,8 @@
 *  in the file PATENTS.  All contributing project authors may
 *  be found in the AUTHORS file in the root of the source tree.
 */
+#include <stdio.h>
+
 #include <climits>
 #include <vector>
 #include "third_party/googletest/src/include/gtest/gtest.h"
@@ -90,34 +92,178 @@ struct FrameInfo {
  unsigned int h;
 };

-unsigned int ScaleForFrameNumber(unsigned int frame, unsigned int val) {
-  if (frame < 10)
-    return val;
-  if (frame < 20)
-    return val / 2;
-  if (frame < 30)
-    return val * 2 / 3;
-  if (frame < 40)
-    return val / 4;
-  if (frame < 50)
-    return val * 7 / 8;
-  return val;
+void ScaleForFrameNumber(unsigned int frame,
+                         unsigned int initial_w,
+                         unsigned int initial_h,
+                         unsigned int *w,
+                         unsigned int *h,
+                         int flag_codec) {
+  if (frame < 10) {
+    *w = initial_w;
+    *h = initial_h;
+    return;
+  }
+  if (frame < 20) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 30) {
+    *w = initial_w / 2;
+    *h = initial_h / 2;
+    return;
+  }
+  if (frame < 40) {
+    *w = initial_w;
+    *h = initial_h;
+    return;
+  }
+  if (frame < 50) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 60) {
+    *w = initial_w / 2;
+    *h = initial_h / 2;
+    return;
+  }
+  if (frame < 70) {
+    *w = initial_w;
+    *h = initial_h;
+    return;
+  }
+  if (frame < 80) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 90) {
+    *w = initial_w / 2;
+    *h = initial_h / 2;
+    return;
+  }
+  if (frame < 100) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 110) {
+    *w = initial_w;
+    *h = initial_h;
+    return;
+  }
+  if (frame < 120) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 130) {
+    *w = initial_w / 2;
+    *h = initial_h / 2;
+    return;
+  }
+  if (frame < 140) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 150) {
+    *w = initial_w;
+    *h = initial_h;
+    return;
+  }
+  if (frame < 160) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 170) {
+    *w = initial_w / 2;
+    *h = initial_h / 2;
+    return;
+  }
+  if (frame < 180) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 190) {
+    *w = initial_w;
+    *h = initial_h;
+    return;
+  }
+  if (frame < 200) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 210) {
+    *w = initial_w / 2;
+    *h = initial_h / 2;
+    return;
+  }
+  if (frame < 220) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 230) {
+    *w = initial_w;
+    *h = initial_h;
+    return;
+  }
+  if (frame < 240) {
+    *w = initial_w * 3 / 4;
+    *h = initial_h * 3 / 4;
+    return;
+  }
+  if (frame < 250) {
+    *w = initial_w  / 2;
+    *h = initial_h / 2;
+    return;
+  }
+  if (frame < 260) {
+    *w = initial_w;
+    *h = initial_h;
+    return;
+  }
+  // Go down very low.
+  if (frame < 270) {
+    *w = initial_w / 4;
+    *h = initial_h / 4;
+    return;
+  }
+  if (flag_codec == 1) {
+    // Cases that only works for VP9.
+    // For VP9: Swap width and height of original.
+    if (frame < 320) {
+      *w = initial_h;
+      *h = initial_w;
+      return;
+    }
+  }
+  *w = initial_w;
+  *h = initial_h;
 }

 class ResizingVideoSource : public ::libvpx_test::DummyVideoSource {
 public:
  ResizingVideoSource() {
    SetSize(kInitialWidth, kInitialHeight);
-    limit_ = 60;
+    limit_ = 350;
  }
-
+  int flag_codec_;
  virtual ~ResizingVideoSource() {}

 protected:
  virtual void Next() {
    ++frame_;
-    SetSize(ScaleForFrameNumber(frame_, kInitialWidth),
-            ScaleForFrameNumber(frame_, kInitialHeight));
+    unsigned int width;
+    unsigned int height;
+    ScaleForFrameNumber(frame_, kInitialWidth, kInitialHeight, &width, &height,
+                        flag_codec_);
+    SetSize(width, height);
    FillFrame();
  }
 };
@@ -144,15 +290,17 @@ class ResizeTest : public ::libvpx_test::EncoderTest,

 TEST_P(ResizeTest, TestExternalResizeWorks) {
  ResizingVideoSource video;
+  video.flag_codec_ = 0;
  cfg_.g_lag_in_frames = 0;
  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));

  for (std::vector<FrameInfo>::const_iterator info = frame_info_list_.begin();
       info != frame_info_list_.end(); ++info) {
    const unsigned int frame = static_cast<unsigned>(info->pts);
-    const unsigned int expected_w = ScaleForFrameNumber(frame, kInitialWidth);
-    const unsigned int expected_h = ScaleForFrameNumber(frame, kInitialHeight);
-
+    unsigned int expected_w;
+    unsigned int expected_h;
+    ScaleForFrameNumber(frame, kInitialWidth, kInitialHeight,
+                        &expected_w, &expected_h, 0);
    EXPECT_EQ(expected_w, info->w)
        << "Frame " << frame << " had unexpected width";
    EXPECT_EQ(expected_h, info->h)
@@ -317,6 +465,17 @@ class ResizeRealtimeTest : public ::libvpx_test::EncoderTest,
    frame_info_list_.push_back(FrameInfo(pts, img.d_w, img.d_h));
  }

+  virtual void MismatchHook(const vpx_image_t *img1,
+                             const vpx_image_t *img2) {
+    double mismatch_psnr = compute_psnr(img1, img2);
+    mismatch_psnr_ += mismatch_psnr;
+    ++mismatch_nframes_;
+  }
+
+  unsigned int GetMismatchFrames() {
+      return mismatch_nframes_;
+  }
+
  void DefaultConfig() {
    cfg_.rc_buf_initial_sz = 500;
    cfg_.rc_buf_optimal_sz = 600;
@@ -342,24 +501,33 @@ class ResizeRealtimeTest : public ::libvpx_test::EncoderTest,
  std::vector< FrameInfo > frame_info_list_;
  int set_cpu_used_;
  bool change_bitrate_;
+  double mismatch_psnr_;
+  int mismatch_nframes_;
 };

 TEST_P(ResizeRealtimeTest, TestExternalResizeWorks) {
  ResizingVideoSource video;
+  video.flag_codec_ = 1;
  DefaultConfig();
+  // Disable internal resize for this test.
+  cfg_.rc_resize_allowed = 0;
  change_bitrate_ = false;
+  mismatch_psnr_ = 0.0;
+  mismatch_nframes_ = 0;
  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));

  for (std::vector<FrameInfo>::const_iterator info = frame_info_list_.begin();
       info != frame_info_list_.end(); ++info) {
    const unsigned int frame = static_cast<unsigned>(info->pts);
-    const unsigned int expected_w = ScaleForFrameNumber(frame, kInitialWidth);
-    const unsigned int expected_h = ScaleForFrameNumber(frame, kInitialHeight);
-
+    unsigned int expected_w;
+    unsigned int expected_h;
+    ScaleForFrameNumber(frame, kInitialWidth, kInitialHeight,
+                        &expected_w, &expected_h, 1);
    EXPECT_EQ(expected_w, info->w)
        << "Frame " << frame << " had unexpected width";
    EXPECT_EQ(expected_h, info->h)
        << "Frame " << frame << " had unexpected height";
+    EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
  }
 }

@@ -373,6 +541,8 @@ TEST_P(ResizeRealtimeTest, TestInternalResizeDown) {
  cfg_.g_w = 352;
  cfg_.g_h = 288;
  change_bitrate_ = false;
+  mismatch_psnr_ = 0.0;
+  mismatch_nframes_ = 0;
  ASSERT_NO_FATAL_FAILURE(RunLoop(&video));

  unsigned int last_w = cfg_.g_w;
@@ -390,8 +560,13 @@ TEST_P(ResizeRealtimeTest, TestInternalResizeDown) {
    }
  }

+#if CONFIG_VP9_DECODER
  // Verify that we get 1 resize down event in this test.
  ASSERT_EQ(1, resize_count) << "Resizing should occur.";
+  EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
+#else
+  printf("Warning: VP9 decoder unavailable, unable to check resize count!\n");
+#endif
 }

 // Verify the dynamic resizer behavior for real time, 1 pass CBR mode.
@@ -404,6 +579,8 @@ TEST_P(ResizeRealtimeTest, TestInternalResizeDownUpChangeBitRate) {
  cfg_.g_w = 352;
  cfg_.g_h = 288;
  change_bitrate_ = true;
+  mismatch_psnr_ = 0.0;
+  mismatch_nframes_ = 0;
  // Disable dropped frames.
  cfg_.rc_dropframe_thresh = 0;
  // Starting bitrate low.
@@ -431,8 +608,13 @@ TEST_P(ResizeRealtimeTest, TestInternalResizeDownUpChangeBitRate) {
    }
  }

+#if CONFIG_VP9_DECODER
  // Verify that we get 2 resize events in this test.
  ASSERT_EQ(resize_count, 2) << "Resizing should occur twice.";
+  EXPECT_EQ(static_cast<unsigned int>(0), GetMismatchFrames());
+#else
+  printf("Warning: VP9 decoder unavailable, unable to check resize count!\n");
+#endif
 }

 vpx_img_fmt_t CspForFrameNumber(int frame) {
--- a/test/sad_test.cc
+++ b/test/sad_test.cc
@@ -689,27 +689,6 @@ INSTANTIATE_TEST_CASE_P(NEON, SADx4Test, ::testing::ValuesIn(x4d_neon_tests));

 //------------------------------------------------------------------------------
 // x86 functions
-#if HAVE_MMX
-const SadMxNParam mmx_tests[] = {
-  make_tuple(16, 16, &vpx_sad16x16_mmx, -1),
-  make_tuple(16, 8, &vpx_sad16x8_mmx, -1),
-  make_tuple(8, 16, &vpx_sad8x16_mmx, -1),
-  make_tuple(8, 8, &vpx_sad8x8_mmx, -1),
-  make_tuple(4, 4, &vpx_sad4x4_mmx, -1),
-};
-INSTANTIATE_TEST_CASE_P(MMX, SADTest, ::testing::ValuesIn(mmx_tests));
-#endif  // HAVE_MMX
-
-#if HAVE_SSE
-#if CONFIG_USE_X86INC
-const SadMxNx4Param x4d_sse_tests[] = {
-  make_tuple(4, 8, &vpx_sad4x8x4d_sse, -1),
-  make_tuple(4, 4, &vpx_sad4x4x4d_sse, -1),
-};
-INSTANTIATE_TEST_CASE_P(SSE, SADx4Test, ::testing::ValuesIn(x4d_sse_tests));
-#endif  // CONFIG_USE_X86INC
-#endif  // HAVE_SSE
-
 #if HAVE_SSE2
 #if CONFIG_USE_X86INC
 const SadMxNParam sse2_tests[] = {
@@ -828,6 +807,8 @@ const SadMxNx4Param x4d_sse2_tests[] = {
  make_tuple(8, 16, &vpx_sad8x16x4d_sse2, -1),
  make_tuple(8, 8, &vpx_sad8x8x4d_sse2, -1),
  make_tuple(8, 4, &vpx_sad8x4x4d_sse2, -1),
+  make_tuple(4, 8, &vpx_sad4x8x4d_sse2, -1),
+  make_tuple(4, 4, &vpx_sad4x4x4d_sse2, -1),
 #if CONFIG_VP9_HIGHBITDEPTH
  make_tuple(64, 64, &vpx_highbd_sad64x64x4d_sse2, 8),
  make_tuple(64, 32, &vpx_highbd_sad64x32x4d_sse2, 8),
--- a/test/simple_encoder.sh
+++ b/test/simple_encoder.sh
@@ -23,7 +23,7 @@ simple_encoder_verify_environment() {
  fi
 }

-# Runs simple_encoder using the codec specified by $1.
+# Runs simple_encoder using the codec specified by $1 with a frame limit of 100.
 simple_encoder() {
  local encoder="${LIBVPX_BIN_PATH}/simple_encoder${VPX_TEST_EXE_SUFFIX}"
  local codec="$1"
@@ -35,7 +35,7 @@ simple_encoder() {
  fi

  eval "${VPX_TEST_PREFIX}" "${encoder}" "${codec}" "${YUV_RAW_INPUT_WIDTH}" \
-      "${YUV_RAW_INPUT_HEIGHT}" "${YUV_RAW_INPUT}" "${output_file}" 9999 \
+      "${YUV_RAW_INPUT_HEIGHT}" "${YUV_RAW_INPUT}" "${output_file}" 9999 0 100 \
      ${devnull}

  [ -e "${output_file}" ] || return 1
@@ -47,16 +47,13 @@ simple_encoder_vp8() {
  fi
 }

-# TODO(tomfinegan): Add a frame limit param to simple_encoder and enable this
-# test. VP9 is just too slow right now: This test takes 4m30s+ on a fast
-# machine.
-DISABLED_simple_encoder_vp9() {
+simple_encoder_vp9() {
  if [ "$(vp9_encode_available)" = "yes" ]; then
    simple_encoder vp9 || return 1
  fi
 }

 simple_encoder_tests="simple_encoder_vp8
-                      DISABLED_simple_encoder_vp9"
+                      simple_encoder_vp9"

 run_tests simple_encoder_verify_environment "${simple_encoder_tests}"
--- a/test/superframe_test.cc
+++ b/test/superframe_test.cc
@@ -17,7 +17,6 @@
 namespace {

 const int kTestMode = 0;
-const int kSuperframeSyntax = 1;

 typedef std::tr1::tuple<libvpx_test::TestMode,int> SuperframeTestParam;

@@ -32,11 +31,9 @@ class SuperframeTest : public ::libvpx_test::EncoderTest,
    InitializeConfig();
    const SuperframeTestParam input = GET_PARAM(1);
    const libvpx_test::TestMode mode = std::tr1::get<kTestMode>(input);
-    const int syntax = std::tr1::get<kSuperframeSyntax>(input);
    SetMode(mode);
    sf_count_ = 0;
    sf_count_max_ = INT_MAX;
-    is_vp10_style_superframe_ = syntax;
  }

  virtual void TearDown() {
@@ -59,8 +56,7 @@ class SuperframeTest : public ::libvpx_test::EncoderTest,
    const uint8_t marker = buffer[pkt->data.frame.sz - 1];
    const int frames = (marker & 0x7) + 1;
    const int mag = ((marker >> 3) & 3) + 1;
-    const unsigned int index_sz =
-        2 + mag * (frames - is_vp10_style_superframe_);
+    const unsigned int index_sz = 2 + mag * frames;
    if ((marker & 0xe0) == 0xc0 &&
        pkt->data.frame.sz >= index_sz &&
        buffer[pkt->data.frame.sz - index_sz] == marker) {
@@ -85,7 +81,6 @@ class SuperframeTest : public ::libvpx_test::EncoderTest,
    return pkt;
  }

-  int is_vp10_style_superframe_;
  int sf_count_;
  int sf_count_max_;
  vpx_codec_cx_pkt_t modified_pkt_;
@@ -106,8 +101,4 @@ TEST_P(SuperframeTest, TestSuperframeIndexIsOptional) {
 VP9_INSTANTIATE_TEST_CASE(SuperframeTest, ::testing::Combine(
    ::testing::Values(::libvpx_test::kTwoPassGood),
    ::testing::Values(0)));
-
-VP10_INSTANTIATE_TEST_CASE(SuperframeTest, ::testing::Combine(
-    ::testing::Values(::libvpx_test::kTwoPassGood),
-    ::testing::Values(CONFIG_MISC_FIXES)));
 }  // namespace
--- a/test/test-data.mk
+++ b/test/test-data.mk
@@ -562,6 +562,8 @@ LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-03-size-226x224.webm
 LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-03-size-226x224.webm.md5
 LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-03-size-226x226.webm
 LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-03-size-226x226.webm.md5
+LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-03-size-352x288.webm
+LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-03-size-352x288.webm.md5
 LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-05-resize.ivf
 LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-05-resize.ivf.md5
 LIBVPX_TEST_DATA-$(CONFIG_VP9_DECODER) += vp90-2-06-bilinear.webm
--- a/test/test-data.sha1
+++ b/test/test-data.sha1
@@ -550,6 +550,8 @@ d17bc08eedfc60c4c23d576a6c964a21bf854d1f *vp90-2-03-size-226x202.webm
 83c6d8f2969b759e10e5c6542baca1265c874c29 *vp90-2-03-size-226x224.webm.md5
 fe0af2ee47b1e5f6a66db369e2d7e9d870b38dce *vp90-2-03-size-226x226.webm
 94ad19b8b699cea105e2ff18f0df2afd7242bcf7 *vp90-2-03-size-226x226.webm.md5
+52bc1dfd3a97b24d922eb8a31d07527891561f2a *vp90-2-03-size-352x288.webm
+3084d6d0a1eec22e85a394422fbc8faae58930a5 *vp90-2-03-size-352x288.webm.md5
 b6524e4084d15b5d0caaa3d3d1368db30cbee69c *vp90-2-03-deltaq.webm
 65f45ec9a55537aac76104818278e0978f94a678 *vp90-2-03-deltaq.webm.md5
 4dbb87494c7f565ffc266c98d17d0d8c7a5c5aba *vp90-2-05-resize.ivf
--- a/test/test.mk
+++ b/test/test.mk
@@ -18,15 +18,17 @@ LIBVPX_TEST_SRCS-yes += video_source.h
 LIBVPX_TEST_SRCS-yes                   += ../md5_utils.h ../md5_utils.c
 LIBVPX_TEST_SRCS-$(CONFIG_DECODERS)    += ivf_video_source.h
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += ../y4minput.h ../y4minput.c
+LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += altref_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += aq_segment_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += datarate_test.cc
+LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += encode_api_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += error_resilience_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += i420_video_source.h
+LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += realtime_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += resize_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += y4m_video_source.h
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += yuv_video_source.h

-LIBVPX_TEST_SRCS-$(CONFIG_VP8_ENCODER) += altref_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP8_ENCODER) += config_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP8_ENCODER) += cq_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP8_ENCODER) += keyframe_test.cc
@@ -44,6 +46,7 @@ LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += frame_size_tests.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += vp9_lossless_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += vp9_end_to_end_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += vp9_ethread_test.cc
+LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += level_test.cc

 LIBVPX_TEST_SRCS-yes                   += decode_test_driver.cc
 LIBVPX_TEST_SRCS-yes                   += decode_test_driver.h
@@ -58,10 +61,10 @@ LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS)    += y4m_test.cc ../y4menc.c ../y4menc.h

 ## WebM Parsing
 ifeq ($(CONFIG_WEBM_IO), yes)
-LIBWEBM_PARSER_SRCS                    += ../third_party/libwebm/mkvparser.cpp
-LIBWEBM_PARSER_SRCS                    += ../third_party/libwebm/mkvreader.cpp
-LIBWEBM_PARSER_SRCS                    += ../third_party/libwebm/mkvparser.hpp
-LIBWEBM_PARSER_SRCS                    += ../third_party/libwebm/mkvreader.hpp
+LIBWEBM_PARSER_SRCS += ../third_party/libwebm/mkvparser/mkvparser.cc
+LIBWEBM_PARSER_SRCS += ../third_party/libwebm/mkvparser/mkvreader.cc
+LIBWEBM_PARSER_SRCS += ../third_party/libwebm/mkvparser/mkvparser.h
+LIBWEBM_PARSER_SRCS += ../third_party/libwebm/mkvparser/mkvreader.h
 LIBVPX_TEST_SRCS-$(CONFIG_DECODERS)    += $(LIBWEBM_PARSER_SRCS)
 LIBVPX_TEST_SRCS-$(CONFIG_DECODERS)    += ../tools_common.h
 LIBVPX_TEST_SRCS-$(CONFIG_DECODERS)    += ../webmdec.cc
@@ -102,6 +105,7 @@ LIBVPX_TEST_SRCS-yes                   += vp8_boolcoder_test.cc
 LIBVPX_TEST_SRCS-yes                   += vp8_fragments_test.cc
 endif

+LIBVPX_TEST_SRCS-$(CONFIG_POSTPROC)    += add_noise_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_POSTPROC)    += pp_filter_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP8_DECODER) += vp8_decrypt_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP8_ENCODER) += quantize_test.cc
@@ -138,10 +142,13 @@ LIBVPX_TEST_SRCS-yes                   += lpf_8_test.cc
 LIBVPX_TEST_SRCS-yes                   += vp9_intrapred_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_DECODER) += vp9_decrypt_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_DECODER) += vp9_thread_test.cc
+LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += avg_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += dct16x16_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += dct32x32_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += fdct4x4_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += fdct8x8_test.cc
+LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += hadamard_test.cc
+LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += minmax_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += variance_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += vp9_error_block_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += vp9_quantize_test.cc
@@ -160,20 +167,8 @@ LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += vp9_arf_freq_test.cc

 endif # VP9

-## VP10
-ifeq ($(CONFIG_VP10),yes)
-
-LIBVPX_TEST_SRCS-yes                    += vp10_inv_txfm_test.cc
-LIBVPX_TEST_SRCS-$(CONFIG_VP10_ENCODER) += vp10_dct_test.cc
-
-endif # VP10
-
 ## Multi-codec / unconditional whitebox tests.

-ifeq ($(findstring yes,$(CONFIG_VP9_ENCODER)$(CONFIG_VP10_ENCODER)),yes)
-LIBVPX_TEST_SRCS-yes += avg_test.cc
-endif
-
 LIBVPX_TEST_SRCS-$(CONFIG_ENCODERS) += sad_test.cc

 TEST_INTRA_PRED_SPEED_SRCS-yes := test_intra_pred_speed.cc
--- a/test/test_intra_pred_speed.cc
+++ b/test/test_intra_pred_speed.cc
@@ -191,14 +191,15 @@ INTRA_PRED_TEST(C, TestIntraPred4, vpx_dc_predictor_4x4_c,
 INTRA_PRED_TEST(SSE2, TestIntraPred4, vpx_dc_predictor_4x4_sse2,
                vpx_dc_left_predictor_4x4_sse2, vpx_dc_top_predictor_4x4_sse2,
                vpx_dc_128_predictor_4x4_sse2, vpx_v_predictor_4x4_sse2,
-                vpx_h_predictor_4x4_sse2, NULL, NULL, NULL, NULL, NULL, NULL,
+                vpx_h_predictor_4x4_sse2, vpx_d45_predictor_4x4_sse2, NULL,
+                NULL, NULL, vpx_d207_predictor_4x4_sse2, NULL,
                vpx_tm_predictor_4x4_sse2)
 #endif  // HAVE_SSE2 && CONFIG_USE_X86INC

 #if HAVE_SSSE3 && CONFIG_USE_X86INC
 INTRA_PRED_TEST(SSSE3, TestIntraPred4, NULL, NULL, NULL, NULL, NULL,
-                NULL, vpx_d45_predictor_4x4_ssse3, NULL, NULL,
-                vpx_d153_predictor_4x4_ssse3, vpx_d207_predictor_4x4_ssse3,
+                NULL, NULL, NULL, NULL,
+                vpx_d153_predictor_4x4_ssse3, NULL,
                vpx_d63_predictor_4x4_ssse3, NULL)
 #endif  // HAVE_SSSE3 && CONFIG_USE_X86INC

@@ -240,13 +241,13 @@ INTRA_PRED_TEST(C, TestIntraPred8, vpx_dc_predictor_8x8_c,
 INTRA_PRED_TEST(SSE2, TestIntraPred8, vpx_dc_predictor_8x8_sse2,
                vpx_dc_left_predictor_8x8_sse2, vpx_dc_top_predictor_8x8_sse2,
                vpx_dc_128_predictor_8x8_sse2, vpx_v_predictor_8x8_sse2,
-                vpx_h_predictor_8x8_sse2, NULL, NULL, NULL, NULL, NULL,
-                NULL, vpx_tm_predictor_8x8_sse2)
+                vpx_h_predictor_8x8_sse2, vpx_d45_predictor_8x8_sse2, NULL,
+                NULL, NULL, NULL, NULL, vpx_tm_predictor_8x8_sse2)
 #endif  // HAVE_SSE2 && CONFIG_USE_X86INC

 #if HAVE_SSSE3 && CONFIG_USE_X86INC
 INTRA_PRED_TEST(SSSE3, TestIntraPred8, NULL, NULL, NULL, NULL, NULL,
-                NULL, vpx_d45_predictor_8x8_ssse3, NULL, NULL,
+                NULL, NULL, NULL, NULL,
                vpx_d153_predictor_8x8_ssse3, vpx_d207_predictor_8x8_ssse3,
                vpx_d63_predictor_8x8_ssse3, NULL)
 #endif  // HAVE_SSSE3 && CONFIG_USE_X86INC
--- a/test/test_vector_test.cc
+++ b/test/test_vector_test.cc
@@ -103,7 +103,7 @@ TEST_P(TestVectorTest, MD5Match) {
  const int mode = std::tr1::get<kDecodeMode>(input);
  libvpx_test::CompressedVideoSource *video = NULL;
  vpx_codec_flags_t flags = 0;
-  vpx_codec_dec_cfg_t cfg = {0};
+  vpx_codec_dec_cfg_t cfg = vpx_codec_dec_cfg_t();
  char str[256];

  if (mode == kFrameParallelMode) {
--- a/test/test_vectors.cc
+++ b/test/test_vectors.cc
@@ -180,7 +180,8 @@ const char *const kVP9TestVectors[] = {
  "vp90-2-03-size-226x198.webm", "vp90-2-03-size-226x200.webm",
  "vp90-2-03-size-226x202.webm", "vp90-2-03-size-226x208.webm",
  "vp90-2-03-size-226x210.webm", "vp90-2-03-size-226x224.webm",
-  "vp90-2-03-size-226x226.webm", "vp90-2-03-deltaq.webm",
+  "vp90-2-03-size-226x226.webm", "vp90-2-03-size-352x288.webm",
+  "vp90-2-03-deltaq.webm",
  "vp90-2-05-resize.ivf", "vp90-2-06-bilinear.webm",
  "vp90-2-07-frame_parallel.webm", "vp90-2-08-tile_1x2_frame_parallel.webm",
  "vp90-2-08-tile_1x2.webm", "vp90-2-08-tile_1x4_frame_parallel.webm",
--- a/test/tile_independence_test.cc
+++ b/test/tile_independence_test.cc
@@ -103,6 +103,4 @@ TEST_P(TileIndependenceTest, MD5Match) {
 }

 VP9_INSTANTIATE_TEST_CASE(TileIndependenceTest, ::testing::Range(0, 2, 1));
-
-VP10_INSTANTIATE_TEST_CASE(TileIndependenceTest, ::testing::Range(0, 2, 1));
 }  // namespace
--- a/test/twopass_encoder.sh
+++ b/test/twopass_encoder.sh
@@ -23,7 +23,8 @@ twopass_encoder_verify_environment() {
  fi
 }

-# Runs twopass_encoder using the codec specified by $1.
+# Runs twopass_encoder using the codec specified by $1 with a frame limit of
+# 100.
 twopass_encoder() {
  local encoder="${LIBVPX_BIN_PATH}/twopass_encoder${VPX_TEST_EXE_SUFFIX}"
  local codec="$1"
@@ -35,7 +36,7 @@ twopass_encoder() {
  fi

  eval "${VPX_TEST_PREFIX}" "${encoder}" "${codec}" "${YUV_RAW_INPUT_WIDTH}" \
-      "${YUV_RAW_INPUT_HEIGHT}" "${YUV_RAW_INPUT}" "${output_file}" \
+      "${YUV_RAW_INPUT_HEIGHT}" "${YUV_RAW_INPUT}" "${output_file}" 100 \
      ${devnull}

  [ -e "${output_file}" ] || return 1
@@ -47,16 +48,13 @@ twopass_encoder_vp8() {
  fi
 }

-# TODO(tomfinegan): Add a frame limit param to twopass_encoder and enable this
-# test. VP9 is just too slow right now: This test takes 31m16s+ on a fast
-# machine.
-DISABLED_twopass_encoder_vp9() {
+twopass_encoder_vp9() {
  if [ "$(vp9_encode_available)" = "yes" ]; then
    twopass_encoder vp9 || return 1
  fi
 }

 twopass_encoder_tests="twopass_encoder_vp8
-                       DISABLED_twopass_encoder_vp9"
+                       twopass_encoder_vp9"

 run_tests twopass_encoder_verify_environment "${twopass_encoder_tests}"
--- a/test/variance_test.cc
+++ b/test/variance_test.cc
@@ -74,6 +74,10 @@ static unsigned int mb_ss_ref(const int16_t *src) {
  return res;
 }

+/* Note:
+ *  Our codebase calculates the "diff" value in the variance algorithm by
+ *  (src - ref).
+ */
 static uint32_t variance_ref(const uint8_t *src, const uint8_t *ref,
                             int l2w, int l2h, int src_stride_coeff,
                             int ref_stride_coeff, uint32_t *sse_ptr,
@@ -87,14 +91,14 @@ static uint32_t variance_ref(const uint8_t *src, const uint8_t *ref,
    for (int x = 0; x < w; x++) {
      int diff;
      if (!use_high_bit_depth_) {
-        diff = ref[w * y * ref_stride_coeff + x] -
-               src[w * y * src_stride_coeff + x];
+        diff = src[w * y * src_stride_coeff + x] -
+               ref[w * y * ref_stride_coeff + x];
        se += diff;
        sse += diff * diff;
 #if CONFIG_VP9_HIGHBITDEPTH
      } else {
-        diff = CONVERT_TO_SHORTPTR(ref)[w * y * ref_stride_coeff + x] -
-               CONVERT_TO_SHORTPTR(src)[w * y * src_stride_coeff + x];
+        diff = CONVERT_TO_SHORTPTR(src)[w * y * src_stride_coeff + x] -
+               CONVERT_TO_SHORTPTR(ref)[w * y * ref_stride_coeff + x];
        se += diff;
        sse += diff * diff;
 #endif  // CONFIG_VP9_HIGHBITDEPTH
@@ -309,15 +313,15 @@ template<typename VarianceFunctionType>
 void VarianceTest<VarianceFunctionType>::RefTest() {
  for (int i = 0; i < 10; ++i) {
    for (int j = 0; j < block_size_; j++) {
-    if (!use_high_bit_depth_) {
-      src_[j] = rnd_.Rand8();
-      ref_[j] = rnd_.Rand8();
+      if (!use_high_bit_depth_) {
+        src_[j] = rnd_.Rand8();
+        ref_[j] = rnd_.Rand8();
 #if CONFIG_VP9_HIGHBITDEPTH
-    } else {
-      CONVERT_TO_SHORTPTR(src_)[j] = rnd_.Rand16() && mask_;
-      CONVERT_TO_SHORTPTR(ref_)[j] = rnd_.Rand16() && mask_;
+      } else {
+        CONVERT_TO_SHORTPTR(src_)[j] = rnd_.Rand16() & mask_;
+        CONVERT_TO_SHORTPTR(ref_)[j] = rnd_.Rand16() & mask_;
 #endif  // CONFIG_VP9_HIGHBITDEPTH
-    }
+      }
    }
    unsigned int sse1, sse2;
    unsigned int var1;
@@ -328,8 +332,10 @@ void VarianceTest<VarianceFunctionType>::RefTest() {
                                           log2height_, stride_coeff,
                                           stride_coeff, &sse2,
                                           use_high_bit_depth_, bit_depth_);
-    EXPECT_EQ(sse1, sse2);
-    EXPECT_EQ(var1, var2);
+    EXPECT_EQ(sse1, sse2)
+        << "Error at test index: " << i;
+    EXPECT_EQ(var1, var2)
+        << "Error at test index: " << i;
  }
 }

@@ -346,8 +352,8 @@ void VarianceTest<VarianceFunctionType>::RefStrideTest() {
        ref_[ref_ind] = rnd_.Rand8();
 #if CONFIG_VP9_HIGHBITDEPTH
      } else {
-        CONVERT_TO_SHORTPTR(src_)[src_ind] = rnd_.Rand16() && mask_;
-        CONVERT_TO_SHORTPTR(ref_)[ref_ind] = rnd_.Rand16() && mask_;
+        CONVERT_TO_SHORTPTR(src_)[src_ind] = rnd_.Rand16() & mask_;
+        CONVERT_TO_SHORTPTR(ref_)[ref_ind] = rnd_.Rand16() & mask_;
 #endif  // CONFIG_VP9_HIGHBITDEPTH
      }
    }
@@ -361,8 +367,10 @@ void VarianceTest<VarianceFunctionType>::RefStrideTest() {
                                           log2height_, src_stride_coeff,
                                           ref_stride_coeff, &sse2,
                                           use_high_bit_depth_, bit_depth_);
-    EXPECT_EQ(sse1, sse2);
-    EXPECT_EQ(var1, var2);
+    EXPECT_EQ(sse1, sse2)
+        << "Error at test index: " << i;
+    EXPECT_EQ(var1, var2)
+        << "Error at test index: " << i;
  }
 }

@@ -968,30 +976,6 @@ INSTANTIATE_TEST_CASE_P(
        make_tuple(2, 2, &vpx_highbd_12_sub_pixel_avg_variance4x4_c, 12)));
 #endif  // CONFIG_VP9_HIGHBITDEPTH

-#if HAVE_MMX
-INSTANTIATE_TEST_CASE_P(MMX, VpxMseTest,
-                        ::testing::Values(make_tuple(4, 4, &vpx_mse16x16_mmx)));
-
-INSTANTIATE_TEST_CASE_P(MMX, SumOfSquaresTest,
-                        ::testing::Values(vpx_get_mb_ss_mmx));
-
-INSTANTIATE_TEST_CASE_P(
-    MMX, VpxVarianceTest,
-    ::testing::Values(make_tuple(4, 4, &vpx_variance16x16_mmx, 0),
-                      make_tuple(4, 3, &vpx_variance16x8_mmx, 0),
-                      make_tuple(3, 4, &vpx_variance8x16_mmx, 0),
-                      make_tuple(3, 3, &vpx_variance8x8_mmx, 0),
-                      make_tuple(2, 2, &vpx_variance4x4_mmx, 0)));
-
-INSTANTIATE_TEST_CASE_P(
-    MMX, VpxSubpelVarianceTest,
-    ::testing::Values(make_tuple(4, 4, &vpx_sub_pixel_variance16x16_mmx, 0),
-                      make_tuple(4, 3, &vpx_sub_pixel_variance16x8_mmx, 0),
-                      make_tuple(3, 4, &vpx_sub_pixel_variance8x16_mmx, 0),
-                      make_tuple(3, 3, &vpx_sub_pixel_variance8x8_mmx, 0),
-                      make_tuple(2, 2, &vpx_sub_pixel_variance4x4_mmx, 0)));
-#endif  // HAVE_MMX
-
 #if HAVE_SSE2
 INSTANTIATE_TEST_CASE_P(SSE2, SumOfSquaresTest,
                        ::testing::Values(vpx_get_mb_ss_sse2));
@@ -1032,8 +1016,8 @@ INSTANTIATE_TEST_CASE_P(
                      make_tuple(3, 4, &vpx_sub_pixel_variance8x16_sse2, 0),
                      make_tuple(3, 3, &vpx_sub_pixel_variance8x8_sse2, 0),
                      make_tuple(3, 2, &vpx_sub_pixel_variance8x4_sse2, 0),
-                      make_tuple(2, 3, &vpx_sub_pixel_variance4x8_sse, 0),
-                      make_tuple(2, 2, &vpx_sub_pixel_variance4x4_sse, 0)));
+                      make_tuple(2, 3, &vpx_sub_pixel_variance4x8_sse2, 0),
+                      make_tuple(2, 2, &vpx_sub_pixel_variance4x4_sse2, 0)));

 INSTANTIATE_TEST_CASE_P(
    SSE2, VpxSubpelAvgVarianceTest,
@@ -1049,8 +1033,8 @@ INSTANTIATE_TEST_CASE_P(
        make_tuple(3, 4, &vpx_sub_pixel_avg_variance8x16_sse2, 0),
        make_tuple(3, 3, &vpx_sub_pixel_avg_variance8x8_sse2, 0),
        make_tuple(3, 2, &vpx_sub_pixel_avg_variance8x4_sse2, 0),
-        make_tuple(2, 3, &vpx_sub_pixel_avg_variance4x8_sse, 0),
-        make_tuple(2, 2, &vpx_sub_pixel_avg_variance4x4_sse, 0)));
+        make_tuple(2, 3, &vpx_sub_pixel_avg_variance4x8_sse2, 0),
+        make_tuple(2, 2, &vpx_sub_pixel_avg_variance4x4_sse2, 0)));
 #endif  // CONFIG_USE_X86INC

 #if CONFIG_VP9_HIGHBITDEPTH
--- a/test/vp10_dct_test.cc
+++ b/test/vp10_dct_test.cc
@@ -1,111 +0,0 @@
-/*
- *  Copyright (c) 2015 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 <math.h>
-#include <stdlib.h>
-#include <new>
-
-#include "third_party/googletest/src/include/gtest/gtest.h"
-#include "test/acm_random.h"
-#include "test/util.h"
-#include "./vpx_config.h"
-#include "vpx_ports/msvc.h"
-
-#undef CONFIG_COEFFICIENT_RANGE_CHECKING
-#define CONFIG_COEFFICIENT_RANGE_CHECKING 1
-#include "vp10/encoder/dct.c"
-
-using libvpx_test::ACMRandom;
-
-namespace {
-void reference_dct_1d(const double *in, double *out, int size) {
-  const double PI = 3.141592653589793238462643383279502884;
-  const double kInvSqrt2 = 0.707106781186547524400844362104;
-  for (int k = 0; k < size; ++k) {
-    out[k] = 0;
-    for (int n = 0; n < size; ++n) {
-      out[k] += in[n] * cos(PI * (2 * n + 1) * k / (2 * size));
-    }
-    if (k == 0)
-      out[k] = out[k] * kInvSqrt2;
-  }
-}
-
-typedef void (*FdctFuncRef)(const double *in, double *out, int size);
-typedef void (*IdctFuncRef)(const double *in, double *out, int size);
-typedef void (*FdctFunc)(const tran_low_t *in, tran_low_t *out);
-typedef void (*IdctFunc)(const tran_low_t *in, tran_low_t *out);
-
-class TransTestBase {
- public:
-  virtual ~TransTestBase() {}
-
- protected:
-  void RunFwdAccuracyCheck() {
-    tran_low_t *input  = new tran_low_t[txfm_size_];
-    tran_low_t *output = new tran_low_t[txfm_size_];
-    double *ref_input  = new double[txfm_size_];
-    double *ref_output = new double[txfm_size_];
-
-    ACMRandom rnd(ACMRandom::DeterministicSeed());
-    const int count_test_block = 5000;
-    for (int ti =  0; ti < count_test_block; ++ti) {
-      for (int ni = 0; ni < txfm_size_; ++ni) {
-        input[ni] = rnd.Rand8() - rnd.Rand8();
-        ref_input[ni] = static_cast<double>(input[ni]);
-      }
-
-      fwd_txfm_(input, output);
-      fwd_txfm_ref_(ref_input, ref_output, txfm_size_);
-
-      for (int ni = 0; ni < txfm_size_; ++ni) {
-        EXPECT_LE(
-            abs(output[ni] - static_cast<tran_low_t>(round(ref_output[ni]))),
-            max_error_);
-      }
-    }
-
-    delete[] input;
-    delete[] output;
-    delete[] ref_input;
-    delete[] ref_output;
-  }
-
-  double max_error_;
-  int txfm_size_;
-  FdctFunc fwd_txfm_;
-  FdctFuncRef fwd_txfm_ref_;
-};
-
-typedef std::tr1::tuple<FdctFunc, FdctFuncRef, int, int> FdctParam;
-class Vp10FwdTxfm
-    : public TransTestBase,
-      public ::testing::TestWithParam<FdctParam> {
- public:
-  virtual void SetUp() {
-    fwd_txfm_ = GET_PARAM(0);
-    fwd_txfm_ref_ = GET_PARAM(1);
-    txfm_size_ = GET_PARAM(2);
-    max_error_ = GET_PARAM(3);
-  }
-  virtual void TearDown() {}
-};
-
-TEST_P(Vp10FwdTxfm, RunFwdAccuracyCheck) {
-  RunFwdAccuracyCheck();
-}
-
-INSTANTIATE_TEST_CASE_P(
-    C, Vp10FwdTxfm,
-    ::testing::Values(
-        FdctParam(&fdct4, &reference_dct_1d, 4, 1),
-        FdctParam(&fdct8, &reference_dct_1d, 8, 1),
-        FdctParam(&fdct16, &reference_dct_1d, 16, 2)));
-}  // namespace
--- a/test/vp10_inv_txfm_test.cc
+++ b/test/vp10_inv_txfm_test.cc
@@ -1,321 +0,0 @@
-/*
- *  Copyright (c) 2013 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 <math.h>
-#include <stdlib.h>
-#include <string.h>
-
-#include "third_party/googletest/src/include/gtest/gtest.h"
-
-#include "./vp10_rtcd.h"
-#include "./vpx_dsp_rtcd.h"
-#include "test/acm_random.h"
-#include "test/clear_system_state.h"
-#include "test/register_state_check.h"
-#include "test/util.h"
-#include "vp10/common/blockd.h"
-#include "vp10/common/scan.h"
-#include "vpx/vpx_integer.h"
-#include "vp10/common/vp10_inv_txfm.h"
-
-using libvpx_test::ACMRandom;
-
-namespace {
-const double PI = 3.141592653589793238462643383279502884;
-const double kInvSqrt2 = 0.707106781186547524400844362104;
-
-void reference_idct_1d(const double *in, double *out, int size) {
-  for (int n = 0; n < size; ++n) {
-    out[n] = 0;
-    for (int k = 0; k < size; ++k) {
-      if (k == 0)
-        out[n] += kInvSqrt2 * in[k] * cos(PI * (2 * n + 1) * k / (2 * size));
-      else
-        out[n] += in[k] * cos(PI * (2 * n + 1) * k / (2 * size));
-    }
-  }
-}
-
-typedef void (*IdctFuncRef)(const double *in, double *out, int size);
-typedef void (*IdctFunc)(const tran_low_t *in, tran_low_t *out);
-
-class TransTestBase {
- public:
-  virtual ~TransTestBase() {}
-
- protected:
-  void RunInvAccuracyCheck() {
-    tran_low_t *input  = new tran_low_t[txfm_size_];
-    tran_low_t *output = new tran_low_t[txfm_size_];
-    double *ref_input  = new double[txfm_size_];
-    double *ref_output = new double[txfm_size_];
-
-    ACMRandom rnd(ACMRandom::DeterministicSeed());
-    const int count_test_block = 5000;
-    for (int ti =  0; ti < count_test_block; ++ti) {
-      for (int ni = 0; ni < txfm_size_; ++ni) {
-        input[ni] = rnd.Rand8() - rnd.Rand8();
-        ref_input[ni] = static_cast<double>(input[ni]);
-      }
-
-      fwd_txfm_(input, output);
-      fwd_txfm_ref_(ref_input, ref_output, txfm_size_);
-
-      for (int ni = 0; ni < txfm_size_; ++ni) {
-        EXPECT_LE(
-            abs(output[ni] - static_cast<tran_low_t>(round(ref_output[ni]))),
-            max_error_);
-      }
-    }
-
-    delete[] input;
-    delete[] output;
-    delete[] ref_input;
-    delete[] ref_output;
-  }
-
-  double max_error_;
-  int txfm_size_;
-  IdctFunc fwd_txfm_;
-  IdctFuncRef fwd_txfm_ref_;
-};
-
-typedef std::tr1::tuple<IdctFunc, IdctFuncRef, int, int> IdctParam;
-class Vp10InvTxfm
-    : public TransTestBase,
-      public ::testing::TestWithParam<IdctParam> {
- public:
-  virtual void SetUp() {
-    fwd_txfm_ = GET_PARAM(0);
-    fwd_txfm_ref_ = GET_PARAM(1);
-    txfm_size_ = GET_PARAM(2);
-    max_error_ = GET_PARAM(3);
-  }
-  virtual void TearDown() {}
-};
-
-TEST_P(Vp10InvTxfm, RunInvAccuracyCheck) {
-  RunInvAccuracyCheck();
-}
-
-INSTANTIATE_TEST_CASE_P(
-    C, Vp10InvTxfm,
-    ::testing::Values(
-        IdctParam(&vp10_idct4_c, &reference_idct_1d, 4, 1),
-        IdctParam(&vp10_idct8_c, &reference_idct_1d, 8, 2),
-        IdctParam(&vp10_idct16_c, &reference_idct_1d, 16, 4),
-        IdctParam(&vp10_idct32_c, &reference_idct_1d, 32, 6))
-);
-
-typedef void (*FwdTxfmFunc)(const int16_t *in, tran_low_t *out, int stride);
-typedef void (*InvTxfmFunc)(const tran_low_t *in, uint8_t *out, int stride);
-typedef std::tr1::tuple<FwdTxfmFunc,
-                        InvTxfmFunc,
-                        InvTxfmFunc,
-                        TX_SIZE, int> PartialInvTxfmParam;
-const int kMaxNumCoeffs = 1024;
-class Vp10PartialIDctTest
-    : public ::testing::TestWithParam<PartialInvTxfmParam> {
- public:
-  virtual ~Vp10PartialIDctTest() {}
-  virtual void SetUp() {
-    ftxfm_ = GET_PARAM(0);
-    full_itxfm_ = GET_PARAM(1);
-    partial_itxfm_ = GET_PARAM(2);
-    tx_size_  = GET_PARAM(3);
-    last_nonzero_ = GET_PARAM(4);
-  }
-
-  virtual void TearDown() { libvpx_test::ClearSystemState(); }
-
- protected:
-  int last_nonzero_;
-  TX_SIZE tx_size_;
-  FwdTxfmFunc ftxfm_;
-  InvTxfmFunc full_itxfm_;
-  InvTxfmFunc partial_itxfm_;
-};
-
-TEST_P(Vp10PartialIDctTest, RunQuantCheck) {
-  ACMRandom rnd(ACMRandom::DeterministicSeed());
-  int size;
-  switch (tx_size_) {
-    case TX_4X4:
-      size = 4;
-      break;
-    case TX_8X8:
-      size = 8;
-      break;
-    case TX_16X16:
-      size = 16;
-      break;
-    case TX_32X32:
-      size = 32;
-      break;
-    default:
-      FAIL() << "Wrong Size!";
-      break;
-  }
-  DECLARE_ALIGNED(16, tran_low_t, test_coef_block1[kMaxNumCoeffs]);
-  DECLARE_ALIGNED(16, tran_low_t, test_coef_block2[kMaxNumCoeffs]);
-  DECLARE_ALIGNED(16, uint8_t, dst1[kMaxNumCoeffs]);
-  DECLARE_ALIGNED(16, uint8_t, dst2[kMaxNumCoeffs]);
-
-  const int count_test_block = 1000;
-  const int block_size = size * size;
-
-  DECLARE_ALIGNED(16, int16_t, input_extreme_block[kMaxNumCoeffs]);
-  DECLARE_ALIGNED(16, tran_low_t, output_ref_block[kMaxNumCoeffs]);
-
-  int max_error = 0;
-  for (int i = 0; i < count_test_block; ++i) {
-    // clear out destination buffer
-    memset(dst1, 0, sizeof(*dst1) * block_size);
-    memset(dst2, 0, sizeof(*dst2) * block_size);
-    memset(test_coef_block1, 0, sizeof(*test_coef_block1) * block_size);
-    memset(test_coef_block2, 0, sizeof(*test_coef_block2) * block_size);
-
-    ACMRandom rnd(ACMRandom::DeterministicSeed());
-
-    for (int i = 0; i < count_test_block; ++i) {
-      // Initialize a test block with input range [-255, 255].
-      if (i == 0) {
-        for (int j = 0; j < block_size; ++j)
-          input_extreme_block[j] = 255;
-      } else if (i == 1) {
-        for (int j = 0; j < block_size; ++j)
-          input_extreme_block[j] = -255;
-      } else {
-        for (int j = 0; j < block_size; ++j) {
-          input_extreme_block[j] = rnd.Rand8() % 2 ? 255 : -255;
-        }
-      }
-
-      ftxfm_(input_extreme_block, output_ref_block, size);
-
-      // quantization with maximum allowed step sizes
-      test_coef_block1[0] = (output_ref_block[0] / 1336) * 1336;
-      for (int j = 1; j < last_nonzero_; ++j)
-        test_coef_block1[vp10_default_scan_orders[tx_size_].scan[j]]
-                         = (output_ref_block[j] / 1828) * 1828;
-    }
-
-    ASM_REGISTER_STATE_CHECK(full_itxfm_(test_coef_block1, dst1, size));
-    ASM_REGISTER_STATE_CHECK(partial_itxfm_(test_coef_block1, dst2, size));
-
-    for (int j = 0; j < block_size; ++j) {
-      const int diff = dst1[j] - dst2[j];
-      const int error = diff * diff;
-      if (max_error < error)
-        max_error = error;
-    }
-  }
-
-  EXPECT_EQ(0, max_error)
-      << "Error: partial inverse transform produces different results";
-}
-
-TEST_P(Vp10PartialIDctTest, ResultsMatch) {
-  ACMRandom rnd(ACMRandom::DeterministicSeed());
-  int size;
-  switch (tx_size_) {
-    case TX_4X4:
-      size = 4;
-      break;
-    case TX_8X8:
-      size = 8;
-      break;
-    case TX_16X16:
-      size = 16;
-      break;
-    case TX_32X32:
-      size = 32;
-      break;
-    default:
-      FAIL() << "Wrong Size!";
-      break;
-  }
-  DECLARE_ALIGNED(16, tran_low_t, test_coef_block1[kMaxNumCoeffs]);
-  DECLARE_ALIGNED(16, tran_low_t, test_coef_block2[kMaxNumCoeffs]);
-  DECLARE_ALIGNED(16, uint8_t, dst1[kMaxNumCoeffs]);
-  DECLARE_ALIGNED(16, uint8_t, dst2[kMaxNumCoeffs]);
-  const int count_test_block = 1000;
-  const int max_coeff = 32766 / 4;
-  const int block_size = size * size;
-  int max_error = 0;
-  for (int i = 0; i < count_test_block; ++i) {
-    // clear out destination buffer
-    memset(dst1, 0, sizeof(*dst1) * block_size);
-    memset(dst2, 0, sizeof(*dst2) * block_size);
-    memset(test_coef_block1, 0, sizeof(*test_coef_block1) * block_size);
-    memset(test_coef_block2, 0, sizeof(*test_coef_block2) * block_size);
-    int max_energy_leftover = max_coeff * max_coeff;
-    for (int j = 0; j < last_nonzero_; ++j) {
-      int16_t coef = static_cast<int16_t>(sqrt(1.0 * max_energy_leftover) *
-                                          (rnd.Rand16() - 32768) / 65536);
-      max_energy_leftover -= coef * coef;
-      if (max_energy_leftover < 0) {
-        max_energy_leftover = 0;
-        coef = 0;
-      }
-      test_coef_block1[vp10_default_scan_orders[tx_size_].scan[j]] = coef;
-    }
-
-    memcpy(test_coef_block2, test_coef_block1,
-           sizeof(*test_coef_block2) * block_size);
-
-    ASM_REGISTER_STATE_CHECK(full_itxfm_(test_coef_block1, dst1, size));
-    ASM_REGISTER_STATE_CHECK(partial_itxfm_(test_coef_block2, dst2, size));
-
-    for (int j = 0; j < block_size; ++j) {
-      const int diff = dst1[j] - dst2[j];
-      const int error = diff * diff;
-      if (max_error < error)
-        max_error = error;
-    }
-  }
-
-  EXPECT_EQ(0, max_error)
-      << "Error: partial inverse transform produces different results";
-}
-using std::tr1::make_tuple;
-
-INSTANTIATE_TEST_CASE_P(
-    C, Vp10PartialIDctTest,
-    ::testing::Values(
-        make_tuple(&vpx_fdct32x32_c,
-                   &vp10_idct32x32_1024_add_c,
-                   &vp10_idct32x32_34_add_c,
-                   TX_32X32, 34),
-        make_tuple(&vpx_fdct32x32_c,
-                   &vp10_idct32x32_1024_add_c,
-                   &vp10_idct32x32_1_add_c,
-                   TX_32X32, 1),
-        make_tuple(&vpx_fdct16x16_c,
-                   &vp10_idct16x16_256_add_c,
-                   &vp10_idct16x16_10_add_c,
-                   TX_16X16, 10),
-        make_tuple(&vpx_fdct16x16_c,
-                   &vp10_idct16x16_256_add_c,
-                   &vp10_idct16x16_1_add_c,
-                   TX_16X16, 1),
-        make_tuple(&vpx_fdct8x8_c,
-                   &vp10_idct8x8_64_add_c,
-                   &vp10_idct8x8_12_add_c,
-                   TX_8X8, 12),
-        make_tuple(&vpx_fdct8x8_c,
-                   &vp10_idct8x8_64_add_c,
-                   &vp10_idct8x8_1_add_c,
-                   TX_8X8, 1),
-        make_tuple(&vpx_fdct4x4_c,
-                   &vp10_idct4x4_16_add_c,
-                   &vp10_idct4x4_1_add_c,
-                   TX_4X4, 1)));
-}  // namespace
--- a/test/vp9_arf_freq_test.cc
+++ b/test/vp9_arf_freq_test.cc
@@ -229,24 +229,4 @@ VP9_INSTANTIATE_TEST_CASE(
    ::testing::ValuesIn(kTestVectors),
    ::testing::ValuesIn(kEncodeVectors),
    ::testing::ValuesIn(kMinArfVectors));
-
-#if CONFIG_VP9_HIGHBITDEPTH
-# if CONFIG_VP10_ENCODER
-// TODO(angiebird): 25-29 fail in high bitdepth mode.
-INSTANTIATE_TEST_CASE_P(
-    DISABLED_VP10, ArfFreqTest,
-    ::testing::Combine(
-        ::testing::Values(static_cast<const libvpx_test::CodecFactory *>(
-            &libvpx_test::kVP10)),
-        ::testing::ValuesIn(kTestVectors),
-        ::testing::ValuesIn(kEncodeVectors),
-        ::testing::ValuesIn(kMinArfVectors)));
-# endif  // CONFIG_VP10_ENCODER
-#else
-VP10_INSTANTIATE_TEST_CASE(
-    ArfFreqTest,
-    ::testing::ValuesIn(kTestVectors),
-    ::testing::ValuesIn(kEncodeVectors),
-    ::testing::ValuesIn(kMinArfVectors));
-#endif  // CONFIG_VP9_HIGHBITDEPTH
 }  // namespace
--- a/test/vp9_denoiser_sse2_test.cc
+++ b/test/vp9_denoiser_sse2_test.cc
@@ -94,8 +94,7 @@ TEST_P(VP9DenoiserTest, BitexactCheck) {
 // Test for all block size.
 INSTANTIATE_TEST_CASE_P(
    SSE2, VP9DenoiserTest,
-    ::testing::Values(BLOCK_4X4, BLOCK_4X8, BLOCK_8X4, BLOCK_8X8,
-                      BLOCK_8X16, BLOCK_16X8, BLOCK_16X16, BLOCK_16X32,
-                      BLOCK_32X16, BLOCK_32X32, BLOCK_32X64, BLOCK_64X32,
-                      BLOCK_64X64));
+    ::testing::Values(BLOCK_8X8, BLOCK_8X16, BLOCK_16X8, BLOCK_16X16,
+                      BLOCK_16X32, BLOCK_32X16, BLOCK_32X32, BLOCK_32X64,
+                      BLOCK_64X32, BLOCK_64X64));
 }  // namespace
--- a/test/vp9_encoder_parms_get_to_decoder.cc
+++ b/test/vp9_encoder_parms_get_to_decoder.cc
@@ -45,9 +45,9 @@ struct EncodeParameters {
 };

 const EncodeParameters kVP9EncodeParameterSet[] = {
-  {0, 0, 0, 1, 0, VPX_CR_STUDIO_RANGE, VPX_CS_BT_601},
-  {0, 0, 0, 0, 0, VPX_CR_FULL_RANGE, VPX_CS_BT_709},
-  {0, 0, 1, 0, 0, VPX_CR_FULL_RANGE, VPX_CS_BT_2020},
+  {0, 0, 0, 1, 0, VPX_CR_STUDIO_RANGE, VPX_CS_BT_601, { 0, 0 }},
+  {0, 0, 0, 0, 0, VPX_CR_FULL_RANGE, VPX_CS_BT_709, { 0, 0 }},
+  {0, 0, 1, 0, 0, VPX_CR_FULL_RANGE, VPX_CS_BT_2020, { 0, 0 }},
  {0, 2, 0, 0, 1, VPX_CR_STUDIO_RANGE, VPX_CS_UNKNOWN, { 640, 480 }},
  // TODO(JBB): Test profiles (requires more work).
 };
@@ -93,7 +93,7 @@ class VpxEncoderParmsGetToDecoder
  }

  virtual bool HandleDecodeResult(const vpx_codec_err_t res_dec,
-                                  const libvpx_test::VideoSource &video,
+                                  const libvpx_test::VideoSource & /*video*/,
                                  libvpx_test::Decoder *decoder) {
    vpx_codec_ctx_t *const vp9_decoder = decoder->GetDecoder();
    vpx_codec_alg_priv_t *const priv =
--- a/test/vp9_end_to_end_test.cc
+++ b/test/vp9_end_to_end_test.cc
@@ -186,24 +186,4 @@ VP9_INSTANTIATE_TEST_CASE(
    ::testing::ValuesIn(kEncodingModeVectors),
    ::testing::ValuesIn(kTestVectors),
    ::testing::ValuesIn(kCpuUsedVectors));
-
-#if CONFIG_VP9_HIGHBITDEPTH
-# if CONFIG_VP10_ENCODER
-// TODO(angiebird): many fail in high bitdepth mode.
-INSTANTIATE_TEST_CASE_P(
-    DISABLED_VP10, EndToEndTestLarge,
-    ::testing::Combine(
-        ::testing::Values(static_cast<const libvpx_test::CodecFactory *>(
-            &libvpx_test::kVP10)),
-        ::testing::ValuesIn(kEncodingModeVectors),
-        ::testing::ValuesIn(kTestVectors),
-        ::testing::ValuesIn(kCpuUsedVectors)));
-# endif  // CONFIG_VP10_ENCODER
-#else
-VP10_INSTANTIATE_TEST_CASE(
-    EndToEndTestLarge,
-    ::testing::ValuesIn(kEncodingModeVectors),
-    ::testing::ValuesIn(kTestVectors),
-    ::testing::ValuesIn(kCpuUsedVectors));
-#endif  // CONFIG_VP9_HIGHBITDEPTH
 }  // namespace
--- a/test/vp9_error_block_test.cc
+++ b/test/vp9_error_block_test.cc
@@ -164,7 +164,7 @@ int64_t wrap_vp9_highbd_block_error_8bit_c(const tran_low_t *coeff,
                                           const tran_low_t *dqcoeff,
                                           intptr_t block_size,
                                           int64_t *ssz, int bps) {
-  assert(bps == 8);
+  EXPECT_EQ(8, bps);
  return vp9_highbd_block_error_8bit_c(coeff, dqcoeff, block_size, ssz);
 }

@@ -173,7 +173,7 @@ int64_t wrap_vp9_highbd_block_error_8bit_sse2(const tran_low_t *coeff,
                                              const tran_low_t *dqcoeff,
                                              intptr_t block_size,
                                              int64_t *ssz, int bps) {
-  assert(bps == 8);
+  EXPECT_EQ(8, bps);
  return vp9_highbd_block_error_8bit_sse2(coeff, dqcoeff, block_size, ssz);
 }

@@ -195,7 +195,7 @@ int64_t wrap_vp9_highbd_block_error_8bit_avx(const tran_low_t *coeff,
                                              const tran_low_t *dqcoeff,
                                              intptr_t block_size,
                                              int64_t *ssz, int bps) {
-  assert(bps == 8);
+  EXPECT_EQ(8, bps);
  return vp9_highbd_block_error_8bit_avx(coeff, dqcoeff, block_size, ssz);
 }

--- a/test/vp9_ethread_test.cc
+++ b/test/vp9_ethread_test.cc
@@ -29,16 +29,9 @@ class VPxEncoderThreadTest
        encoding_mode_(GET_PARAM(1)),
        set_cpu_used_(GET_PARAM(2)) {
    init_flags_ = VPX_CODEC_USE_PSNR;
-    vpx_codec_dec_cfg_t cfg = vpx_codec_dec_cfg_t();
-    cfg.w = 1280;
-    cfg.h = 720;
-    decoder_ = codec_->CreateDecoder(cfg, 0);
-
    md5_.clear();
  }
-  virtual ~VPxEncoderThreadTest() {
-    delete decoder_;
-  }
+  virtual ~VPxEncoderThreadTest() {}

  virtual void SetUp() {
    InitializeConfig();
@@ -48,7 +41,7 @@ class VPxEncoderThreadTest
      cfg_.g_lag_in_frames = 3;
      cfg_.rc_end_usage = VPX_VBR;
      cfg_.rc_2pass_vbr_minsection_pct = 5;
-      cfg_.rc_2pass_vbr_minsection_pct = 2000;
+      cfg_.rc_2pass_vbr_maxsection_pct = 2000;
    } else {
      cfg_.g_lag_in_frames = 0;
      cfg_.rc_end_usage = VPX_CBR;
@@ -62,7 +55,7 @@ class VPxEncoderThreadTest
    encoder_initialized_ = false;
  }

-  virtual void PreEncodeFrameHook(::libvpx_test::VideoSource *video,
+  virtual void PreEncodeFrameHook(::libvpx_test::VideoSource * /*video*/,
                                  ::libvpx_test::Encoder *encoder) {
    if (!encoder_initialized_) {
      // Encode 4 column tiles.
@@ -81,27 +74,28 @@ class VPxEncoderThreadTest
    }
  }

-  virtual void FramePktHook(const vpx_codec_cx_pkt_t *pkt) {
-    const vpx_codec_err_t res = decoder_->DecodeFrame(
-        reinterpret_cast<uint8_t*>(pkt->data.frame.buf), pkt->data.frame.sz);
-    if (res != VPX_CODEC_OK) {
-      abort_ = true;
-      ASSERT_EQ(VPX_CODEC_OK, res);
-    }
-    const vpx_image_t *img = decoder_->GetDxData().Next();
+  virtual void DecompressedFrameHook(const vpx_image_t &img,
+                                     vpx_codec_pts_t /*pts*/) {
+    ::libvpx_test::MD5 md5_res;
+    md5_res.Add(&img);
+    md5_.push_back(md5_res.Get());
+  }

-    if (img) {
-      ::libvpx_test::MD5 md5_res;
-      md5_res.Add(img);
-      md5_.push_back(md5_res.Get());
+  virtual bool HandleDecodeResult(const vpx_codec_err_t res,
+                                  const libvpx_test::VideoSource& /*video*/,
+                                  libvpx_test::Decoder * /*decoder*/) {
+    if (res != VPX_CODEC_OK) {
+      EXPECT_EQ(VPX_CODEC_OK, res);
+      return false;
    }
+
+    return true;
  }

  bool encoder_initialized_;
  int tiles_;
  ::libvpx_test::TestMode encoding_mode_;
  int set_cpu_used_;
-  ::libvpx_test::Decoder *decoder_;
  std::vector<std::string> md5_;
 };

@@ -134,9 +128,4 @@ VP9_INSTANTIATE_TEST_CASE(
    ::testing::Values(::libvpx_test::kTwoPassGood, ::libvpx_test::kOnePassGood,
                      ::libvpx_test::kRealTime),
    ::testing::Range(1, 9));
-
-VP10_INSTANTIATE_TEST_CASE(
-    VPxEncoderThreadTest,
-    ::testing::Values(::libvpx_test::kTwoPassGood, ::libvpx_test::kOnePassGood),
-    ::testing::Range(1, 3));
 }  // namespace
--- a/test/vp9_intrapred_test.cc
+++ b/test/vp9_intrapred_test.cc
@@ -34,7 +34,7 @@ class VP9IntraPredBase {
  virtual ~VP9IntraPredBase() { libvpx_test::ClearSystemState(); }

 protected:
-  virtual void Predict(PREDICTION_MODE mode) = 0;
+  virtual void Predict() = 0;

  void CheckPrediction(int test_case_number, int *error_count) const {
    // For each pixel ensure that the calculated value is the same as reference.
@@ -73,7 +73,7 @@ class VP9IntraPredBase {
          left_col_[y] = rnd.Rand16() & mask_;
        }
      }
-      Predict(DC_PRED);
+      Predict();
      CheckPrediction(i, &error_count);
    }
    ASSERT_EQ(0, error_count);
@@ -106,7 +106,7 @@ class VP9IntraPredTest
    mask_       = (1 << bit_depth_) - 1;
  }

-  virtual void Predict(PREDICTION_MODE mode) {
+  virtual void Predict() {
    const uint16_t *const_above_row = above_row_;
    const uint16_t *const_left_col = left_col_;
    ref_fn_(ref_dst_, stride_, const_above_row, const_left_col, bit_depth_);
--- a/test/vp9_lossless_test.cc
+++ b/test/vp9_lossless_test.cc
@@ -127,8 +127,4 @@ VP9_INSTANTIATE_TEST_CASE(LosslessTest,
                          ::testing::Values(::libvpx_test::kRealTime,
                                            ::libvpx_test::kOnePassGood,
                                            ::libvpx_test::kTwoPassGood));
-
-VP10_INSTANTIATE_TEST_CASE(LosslessTest,
-                           ::testing::Values(::libvpx_test::kOnePassGood,
-                                             ::libvpx_test::kTwoPassGood));
 }  // namespace
--- a/test/webm_video_source.h
+++ b/test/webm_video_source.h
@@ -62,7 +62,7 @@ class WebMVideoSource : public CompressedVideoSource {

  void FillFrame() {
    ASSERT_TRUE(vpx_ctx_->file != NULL);
-    const int status = webm_read_frame(webm_ctx_, &buf_, &buf_sz_, &buf_sz_);
+    const int status = webm_read_frame(webm_ctx_, &buf_, &buf_sz_);
    ASSERT_GE(status, 0) << "webm_read_frame failed";
    if (status == 1) {
      end_of_file_ = true;
@@ -72,7 +72,7 @@ class WebMVideoSource : public CompressedVideoSource {
  void SeekToNextKeyFrame() {
    ASSERT_TRUE(vpx_ctx_->file != NULL);
    do {
-      const int status = webm_read_frame(webm_ctx_, &buf_, &buf_sz_, &buf_sz_);
+      const int status = webm_read_frame(webm_ctx_, &buf_, &buf_sz_);
      ASSERT_GE(status, 0) << "webm_read_frame failed";
      ++frame_;
      if (status == 1) {
--- a/third_party/googletest/README.libvpx
+++ b/third_party/googletest/README.libvpx
@@ -12,4 +12,8 @@ failures, various options for running the tests, and XML test report
 generation.

 Local Modifications:
-Removed unused declarations of kPathSeparatorString to have warning free build.
+- Removed unused declarations of kPathSeparatorString to have warning
+  free build.
+- Added GTEST_ATTRIBUTE_UNUSED_ to test registering dummies in TEST_P
+  and INSTANTIATE_TEST_CASE_P to remove warnings about unused variables
+  under GCC 5.
--- a/third_party/googletest/src/include/gtest/gtest.h
+++ b/third_party/googletest/src/include/gtest/gtest.h
@@ -16960,7 +16960,7 @@ internal::CartesianProductHolder10<Generator1, Generator2, Generator3,
                      GTEST_TEST_CLASS_NAME_(test_case_name, test_name)>()); \
      return 0; \
    } \
-    static int gtest_registering_dummy_; \
+    static int gtest_registering_dummy_ GTEST_ATTRIBUTE_UNUSED_; \
    GTEST_DISALLOW_COPY_AND_ASSIGN_(\
        GTEST_TEST_CLASS_NAME_(test_case_name, test_name)); \
  }; \
@@ -16972,7 +16972,7 @@ internal::CartesianProductHolder10<Generator1, Generator2, Generator3,
 # define INSTANTIATE_TEST_CASE_P(prefix, test_case_name, generator) \
  ::testing::internal::ParamGenerator<test_case_name::ParamType> \
      gtest_##prefix##test_case_name##_EvalGenerator_() { return generator; } \
-  int gtest_##prefix##test_case_name##_dummy_ = \
+  int gtest_##prefix##test_case_name##_dummy_ GTEST_ATTRIBUTE_UNUSED_ = \
      ::testing::UnitTest::GetInstance()->parameterized_test_registry(). \
          GetTestCasePatternHolder<test_case_name>(\
              #test_case_name, __FILE__, __LINE__)->AddTestCaseInstantiation(\
--- a/third_party/libwebm/Android.mk
+++ b/third_party/libwebm/Android.mk
@@ -2,9 +2,16 @@ LOCAL_PATH:= $(call my-dir)

 include $(CLEAR_VARS)
 LOCAL_MODULE:= libwebm
-LOCAL_SRC_FILES:= mkvparser.cpp \
-                  mkvreader.cpp \
-                  mkvmuxer.cpp \
-                  mkvmuxerutil.cpp \
-                  mkvwriter.cpp
+LOCAL_CPPFLAGS:=-D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS
+LOCAL_CPPFLAGS+=-D__STDC_LIMIT_MACROS -Wno-extern-c-compat
+LOCAL_C_INCLUDES:= $(LOCAL_PATH)
+LOCAL_EXPORT_C_INCLUDES:= $(LOCAL_PATH)
+
+LOCAL_SRC_FILES:= common/file_util.cc \
+                  common/hdr_util.cc \
+                  mkvparser/mkvparser.cc \
+                  mkvparser/mkvreader.cc \
+                  mkvmuxer/mkvmuxer.cc \
+                  mkvmuxer/mkvmuxerutil.cc \
+                  mkvmuxer/mkvwriter.cc
 include $(BUILD_STATIC_LIBRARY)
--- a/third_party/libwebm/README.libvpx
+++ b/third_party/libwebm/README.libvpx
@@ -1,5 +1,5 @@
 URL: https://chromium.googlesource.com/webm/libwebm
-Version: 476366249e1fda7710a389cd41c57db42305e0d4
+Version: 32d5ac49414a8914ec1e1f285f3f927c6e8ec29d
 License: BSD
 License File: LICENSE.txt

--- a/third_party/libwebm/RELEASE.TXT
+++ b/third_party/libwebm/RELEASE.TXT
@@ -1,34 +0,0 @@
-1.0.0.5
- * Handled case when no duration
- * Handled empty clusters
- * Handled empty clusters when seeking
- * Implemented check lacing bits
-
-1.0.0.4
- * Made Cues member variables mutables
- * Defined against badly-formatted cue points
- * Segment::GetCluster returns CuePoint too
- * Separated cue-based searches
-
-1.0.0.3
- * Added Block::GetOffset() to get a frame's offset in a block
- * Changed cluster count type from size_t to long
- * Parsed SeekHead to find cues
- * Allowed seeking beyond end of cluster cache
- * Added not to attempt to reparse cues element
- * Restructured Segment::LoadCluster
- * Marked position of cues without parsing cues element
- * Allowed cue points to be loaded incrementally
- * Implemented to load lazily cue points as they're searched
- * Merged Cues::LoadCuePoint into Cues::Find
- * Lazy init cues
- * Loaded cue point during find
-
-1.0.0.2
- * added support for Cues element
- * seeking was improved
-
-1.0.0.1
- * fixed item 141
- * added item 142
- * added this file, RELEASE.TXT, to repository
--- a/third_party/libwebm/common/file_util.cc
+++ b/third_party/libwebm/common/file_util.cc
@@ -0,0 +1,67 @@
+// Copyright (c) 2016 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 "common/file_util.h"
+
+#include <sys/stat.h>
+#ifndef _MSC_VER
+#include <unistd.h>  // close()
+#endif
+
+#include <cstdio>
+#include <cstdlib>
+#include <fstream>
+#include <ios>
+
+namespace libwebm {
+
+std::string GetTempFileName() {
+#if !defined _MSC_VER && !defined __MINGW32__
+  char temp_file_name_template[] = "libwebm_temp.XXXXXX";
+  int fd = mkstemp(temp_file_name_template);
+  if (fd != -1) {
+    close(fd);
+    return std::string(temp_file_name_template);
+  }
+  return std::string();
+#else
+  char tmp_file_name[_MAX_PATH];
+  errno_t err = tmpnam_s(tmp_file_name);
+  if (err == 0) {
+    return std::string(tmp_file_name);
+  }
+  return std::string();
+#endif
+}
+
+uint64_t GetFileSize(const std::string& file_name) {
+  uint64_t file_size = 0;
+#ifndef _MSC_VER
+  struct stat st;
+  st.st_size = 0;
+  if (stat(file_name.c_str(), &st) == 0) {
+#else
+  struct _stat st;
+  st.st_size = 0;
+  if (_stat(file_name.c_str(), &st) == 0) {
+#endif
+    file_size = st.st_size;
+  }
+  return file_size;
+}
+
+TempFileDeleter::TempFileDeleter() { file_name_ = GetTempFileName(); }
+
+TempFileDeleter::~TempFileDeleter() {
+  std::ifstream file(file_name_.c_str());
+  if (file.good()) {
+    file.close();
+    std::remove(file_name_.c_str());
+  }
+}
+
+}  // namespace libwebm
--- a/third_party/libwebm/common/file_util.h
+++ b/third_party/libwebm/common/file_util.h
@@ -0,0 +1,41 @@
+// Copyright (c) 2016 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 LIBWEBM_COMMON_FILE_UTIL_H_
+#define LIBWEBM_COMMON_FILE_UTIL_H_
+
+#include <stdint.h>
+
+#include <string>
+
+#include "mkvmuxer/mkvmuxertypes.h"  // LIBWEBM_DISALLOW_COPY_AND_ASSIGN()
+
+namespace libwebm {
+
+// Returns a temporary file name.
+std::string GetTempFileName();
+
+// Returns size of file specified by |file_name|, or 0 upon failure.
+uint64_t GetFileSize(const std::string& file_name);
+
+// Manages life of temporary file specified at time of construction. Deletes
+// file upon destruction.
+class TempFileDeleter {
+ public:
+  TempFileDeleter();
+  explicit TempFileDeleter(std::string file_name) : file_name_(file_name) {}
+  ~TempFileDeleter();
+  const std::string& name() const { return file_name_; }
+
+ private:
+  std::string file_name_;
+  LIBWEBM_DISALLOW_COPY_AND_ASSIGN(TempFileDeleter);
+};
+
+}  // namespace libwebm
+
+#endif  // LIBWEBM_COMMON_FILE_UTIL_H_
--- a/third_party/libwebm/common/hdr_util.cc
+++ b/third_party/libwebm/common/hdr_util.cc
@@ -0,0 +1,182 @@
+// Copyright (c) 2016 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 "hdr_util.h"
+
+#include <cstddef>
+#include <new>
+
+#include "mkvparser/mkvparser.h"
+
+namespace libwebm {
+bool CopyPrimaryChromaticity(const mkvparser::PrimaryChromaticity& parser_pc,
+                             PrimaryChromaticityPtr* muxer_pc) {
+  muxer_pc->reset(new (std::nothrow)
+                      mkvmuxer::PrimaryChromaticity(parser_pc.x, parser_pc.y));
+  if (!muxer_pc->get())
+    return false;
+  return true;
+}
+
+bool MasteringMetadataValuePresent(double value) {
+  return value != mkvparser::MasteringMetadata::kValueNotPresent;
+}
+
+bool CopyMasteringMetadata(const mkvparser::MasteringMetadata& parser_mm,
+                           mkvmuxer::MasteringMetadata* muxer_mm) {
+  if (MasteringMetadataValuePresent(parser_mm.luminance_max))
+    muxer_mm->luminance_max = parser_mm.luminance_max;
+  if (MasteringMetadataValuePresent(parser_mm.luminance_min))
+    muxer_mm->luminance_min = parser_mm.luminance_min;
+
+  PrimaryChromaticityPtr r_ptr(NULL);
+  PrimaryChromaticityPtr g_ptr(NULL);
+  PrimaryChromaticityPtr b_ptr(NULL);
+  PrimaryChromaticityPtr wp_ptr(NULL);
+
+  if (parser_mm.r) {
+    if (!CopyPrimaryChromaticity(*parser_mm.r, &r_ptr))
+      return false;
+  }
+  if (parser_mm.g) {
+    if (!CopyPrimaryChromaticity(*parser_mm.g, &g_ptr))
+      return false;
+  }
+  if (parser_mm.b) {
+    if (!CopyPrimaryChromaticity(*parser_mm.b, &b_ptr))
+      return false;
+  }
+  if (parser_mm.white_point) {
+    if (!CopyPrimaryChromaticity(*parser_mm.white_point, &wp_ptr))
+      return false;
+  }
+
+  if (!muxer_mm->SetChromaticity(r_ptr.get(), g_ptr.get(), b_ptr.get(),
+                                 wp_ptr.get())) {
+    return false;
+  }
+
+  return true;
+}
+
+bool ColourValuePresent(long long value) {
+  return value != mkvparser::Colour::kValueNotPresent;
+}
+
+bool CopyColour(const mkvparser::Colour& parser_colour,
+                mkvmuxer::Colour* muxer_colour) {
+  if (!muxer_colour)
+    return false;
+
+  if (ColourValuePresent(parser_colour.matrix_coefficients))
+    muxer_colour->matrix_coefficients = parser_colour.matrix_coefficients;
+  if (ColourValuePresent(parser_colour.bits_per_channel))
+    muxer_colour->bits_per_channel = parser_colour.bits_per_channel;
+  if (ColourValuePresent(parser_colour.chroma_subsampling_horz))
+    muxer_colour->chroma_subsampling_horz =
+        parser_colour.chroma_subsampling_horz;
+  if (ColourValuePresent(parser_colour.chroma_subsampling_vert))
+    muxer_colour->chroma_subsampling_vert =
+        parser_colour.chroma_subsampling_vert;
+  if (ColourValuePresent(parser_colour.cb_subsampling_horz))
+    muxer_colour->cb_subsampling_horz = parser_colour.cb_subsampling_horz;
+  if (ColourValuePresent(parser_colour.cb_subsampling_vert))
+    muxer_colour->cb_subsampling_vert = parser_colour.cb_subsampling_vert;
+  if (ColourValuePresent(parser_colour.chroma_siting_horz))
+    muxer_colour->chroma_siting_horz = parser_colour.chroma_siting_horz;
+  if (ColourValuePresent(parser_colour.chroma_siting_vert))
+    muxer_colour->chroma_siting_vert = parser_colour.chroma_siting_vert;
+  if (ColourValuePresent(parser_colour.range))
+    muxer_colour->range = parser_colour.range;
+  if (ColourValuePresent(parser_colour.transfer_characteristics))
+    muxer_colour->transfer_characteristics =
+        parser_colour.transfer_characteristics;
+  if (ColourValuePresent(parser_colour.primaries))
+    muxer_colour->primaries = parser_colour.primaries;
+  if (ColourValuePresent(parser_colour.max_cll))
+    muxer_colour->max_cll = parser_colour.max_cll;
+  if (ColourValuePresent(parser_colour.max_fall))
+    muxer_colour->max_fall = parser_colour.max_fall;
+
+  if (parser_colour.mastering_metadata) {
+    mkvmuxer::MasteringMetadata muxer_mm;
+    if (!CopyMasteringMetadata(*parser_colour.mastering_metadata, &muxer_mm))
+      return false;
+    if (!muxer_colour->SetMasteringMetadata(muxer_mm))
+      return false;
+  }
+  return true;
+}
+
+// Format of VPx private data:
+//
+//   0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+//  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+//  |    ID Byte    |             Length            |               |
+//  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+               |
+//  |                                                               |
+//  :               Bytes 1..Length of Codec Feature                :
+//  |                                                               |
+//  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+//
+// ID Byte Format
+// ID byte is an unsigned byte.
+//   0 1 2 3 4 5 6 7
+//  +-+-+-+-+-+-+-+-+
+//  |X|    ID       |
+//  +-+-+-+-+-+-+-+-+
+//
+// The X bit is reserved.
+//
+// Currently only profile level is supported. ID byte must be set to 1, and
+// length must be 1. Supported values are:
+//
+//   10: Level 1
+//   11: Level 1.1
+//   20: Level 2
+//   21: Level 2.1
+//   30: Level 3
+//   31: Level 3.1
+//   40: Level 4
+//   41: Level 4.1
+//   50: Level 5
+//   51: Level 5.1
+//   52: Level 5.2
+//   60: Level 6
+//   61: Level 6.1
+//   62: Level 6.2
+//
+// See the following link for more information:
+// http://www.webmproject.org/vp9/profiles/
+int ParseVpxCodecPrivate(const uint8_t* private_data, int32_t length) {
+  const int kVpxCodecPrivateLength = 3;
+  if (!private_data || length != kVpxCodecPrivateLength)
+    return 0;
+
+  const uint8_t id_byte = *private_data;
+  if (id_byte != 1)
+    return 0;
+
+  const int kVpxProfileLength = 1;
+  const uint8_t length_byte = private_data[1];
+  if (length_byte != kVpxProfileLength)
+    return 0;
+
+  const int level = static_cast<int>(private_data[2]);
+
+  const int kNumLevels = 14;
+  const int levels[kNumLevels] = {10, 11, 20, 21, 30, 31, 40,
+                                  41, 50, 51, 52, 60, 61, 62};
+
+  for (int i = 0; i < kNumLevels; ++i) {
+    if (level == levels[i])
+      return level;
+  }
+
+  return 0;
+}
+}  // namespace libwebm
--- a/third_party/libwebm/common/hdr_util.h
+++ b/third_party/libwebm/common/hdr_util.h
@@ -0,0 +1,51 @@
+// Copyright (c) 2016 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 LIBWEBM_COMMON_HDR_UTIL_H_
+#define LIBWEBM_COMMON_HDR_UTIL_H_
+
+#include <stdint.h>
+
+#include <memory>
+
+#include "mkvmuxer/mkvmuxer.h"
+
+namespace mkvparser {
+struct Colour;
+struct MasteringMetadata;
+struct PrimaryChromaticity;
+}  // namespace mkvparser
+
+namespace libwebm {
+// Utility types and functions for working with the Colour element and its
+// children. Copiers return true upon success. Presence functions return true
+// when the specified element is present.
+
+// TODO(tomfinegan): These should be moved to libwebm_utils once c++11 is
+// required by libwebm.
+
+typedef std::auto_ptr<mkvmuxer::PrimaryChromaticity> PrimaryChromaticityPtr;
+
+bool CopyPrimaryChromaticity(const mkvparser::PrimaryChromaticity& parser_pc,
+                             PrimaryChromaticityPtr* muxer_pc);
+
+bool MasteringMetadataValuePresent(double value);
+
+bool CopyMasteringMetadata(const mkvparser::MasteringMetadata& parser_mm,
+                           mkvmuxer::MasteringMetadata* muxer_mm);
+
+bool ColourValuePresent(long long value);
+
+bool CopyColour(const mkvparser::Colour& parser_colour,
+                mkvmuxer::Colour* muxer_colour);
+
+// Returns VP9 profile upon success or 0 upon failure.
+int ParseVpxCodecPrivate(const uint8_t* private_data, int32_t length);
+
+}  // namespace libwebm
+
+#endif  // LIBWEBM_COMMON_HDR_UTIL_H_
--- a/third_party/libwebm/common/webmids.h
+++ b/third_party/libwebm/common/webmids.h
@@ -6,10 +6,10 @@
 // in the file PATENTS.  All contributing project authors may
 // be found in the AUTHORS file in the root of the source tree.

-#ifndef WEBMIDS_HPP
-#define WEBMIDS_HPP
+#ifndef COMMON_WEBMIDS_H_
+#define COMMON_WEBMIDS_H_

-namespace mkvmuxer {
+namespace libwebm {

 enum MkvId {
  kMkvEBML = 0x1A45DFA3,
@@ -95,6 +95,35 @@ enum MkvId {
  kMkvAspectRatioType = 0x54B3,
  kMkvFrameRate = 0x2383E3,
  // end video
+  // colour
+  kMkvColour = 0x55B0,
+  kMkvMatrixCoefficients = 0x55B1,
+  kMkvBitsPerChannel = 0x55B2,
+  kMkvChromaSubsamplingHorz = 0x55B3,
+  kMkvChromaSubsamplingVert = 0x55B4,
+  kMkvCbSubsamplingHorz = 0x55B5,
+  kMkvCbSubsamplingVert = 0x55B6,
+  kMkvChromaSitingHorz = 0x55B7,
+  kMkvChromaSitingVert = 0x55B8,
+  kMkvRange = 0x55B9,
+  kMkvTransferCharacteristics = 0x55BA,
+  kMkvPrimaries = 0x55BB,
+  kMkvMaxCLL = 0x55BC,
+  kMkvMaxFALL = 0x55BD,
+  // mastering metadata
+  kMkvMasteringMetadata = 0x55D0,
+  kMkvPrimaryRChromaticityX = 0x55D1,
+  kMkvPrimaryRChromaticityY = 0x55D2,
+  kMkvPrimaryGChromaticityX = 0x55D3,
+  kMkvPrimaryGChromaticityY = 0x55D4,
+  kMkvPrimaryBChromaticityX = 0x55D5,
+  kMkvPrimaryBChromaticityY = 0x55D6,
+  kMkvWhitePointChromaticityX = 0x55D7,
+  kMkvWhitePointChromaticityY = 0x55D8,
+  kMkvLuminanceMax = 0x55D9,
+  kMkvLuminanceMin = 0x55DA,
+  // end mastering metadata
+  // end colour
  // audio
  kMkvAudio = 0xE1,
  kMkvSamplingFrequency = 0xB5,
@@ -150,6 +179,6 @@ enum MkvId {
  kMkvTagString = 0x4487
 };

-}  // end namespace mkvmuxer
+}  // namespace libwebm

-#endif  // WEBMIDS_HPP
+#endif  // COMMON_WEBMIDS_H_
--- a/third_party/libwebm/mkvmuxer/mkvmuxer.cc
+++ b/third_party/libwebm/mkvmuxer/mkvmuxer.cc
--- a/third_party/libwebm/mkvmuxer/mkvmuxer.h
+++ b/third_party/libwebm/mkvmuxer/mkvmuxer.h
--- a/third_party/libwebm/mkvmuxer/mkvmuxertypes.h
+++ b/third_party/libwebm/mkvmuxer/mkvmuxertypes.h
@@ -6,8 +6,17 @@
 // in the file PATENTS.  All contributing project authors may
 // be found in the AUTHORS file in the root of the source tree.

-#ifndef MKVMUXERTYPES_HPP
-#define MKVMUXERTYPES_HPP
+#ifndef MKVMUXER_MKVMUXERTYPES_H_
+#define MKVMUXER_MKVMUXERTYPES_H_
+
+namespace mkvmuxer {
+typedef unsigned char uint8;
+typedef short int16;
+typedef int int32;
+typedef unsigned int uint32;
+typedef long long int64;
+typedef unsigned long long uint64;
+}  // namespace mkvmuxer

 // Copied from Chromium basictypes.h
 // A macro to disallow the copy constructor and operator= functions
@@ -16,15 +25,4 @@
  TypeName(const TypeName&);                       \
  void operator=(const TypeName&)

-namespace mkvmuxer {
-
-typedef unsigned char uint8;
-typedef short int16;
-typedef int int32;
-typedef unsigned int uint32;
-typedef long long int64;
-typedef unsigned long long uint64;
-
-}  // end namespace mkvmuxer
-
-#endif  // MKVMUXERTYPES_HPP
+#endif  // MKVMUXER_MKVMUXERTYPES_HPP_
--- a/third_party/libwebm/mkvmuxer/mkvmuxerutil.cc
+++ b/third_party/libwebm/mkvmuxer/mkvmuxerutil.cc
@@ -6,7 +6,7 @@
 // in the file PATENTS.  All contributing project authors may
 // be found in the AUTHORS file in the root of the source tree.

-#include "mkvmuxerutil.hpp"
+#include "mkvmuxer/mkvmuxerutil.h"

 #ifdef __ANDROID__
 #include <fcntl.h>
@@ -20,13 +20,9 @@
 #include <ctime>
 #include <new>

-#include "mkvwriter.hpp"
-#include "webmids.hpp"
-
-#ifdef _MSC_VER
-// Disable MSVC warnings that suggest making code non-portable.
-#pragma warning(disable : 4996)
-#endif
+#include "common/webmids.h"
+#include "mkvmuxer/mkvmuxer.h"
+#include "mkvmuxer/mkvwriter.h"

 namespace mkvmuxer {

@@ -35,64 +31,68 @@ namespace {
 // Date elements are always 8 octets in size.
 const int kDateElementSize = 8;

-uint64 WriteBlock(IMkvWriter* writer, const Frame* const frame, int64 timecode,
-                  uint64 timecode_scale) {
-  uint64 block_additional_elem_size = 0;
-  uint64 block_addid_elem_size = 0;
-  uint64 block_more_payload_size = 0;
-  uint64 block_more_elem_size = 0;
-  uint64 block_additions_payload_size = 0;
-  uint64 block_additions_elem_size = 0;
+uint64_t WriteBlock(IMkvWriter* writer, const Frame* const frame,
+                    int64_t timecode, uint64_t timecode_scale) {
+  uint64_t block_additional_elem_size = 0;
+  uint64_t block_addid_elem_size = 0;
+  uint64_t block_more_payload_size = 0;
+  uint64_t block_more_elem_size = 0;
+  uint64_t block_additions_payload_size = 0;
+  uint64_t block_additions_elem_size = 0;
  if (frame->additional()) {
-    block_additional_elem_size = EbmlElementSize(
-        kMkvBlockAdditional, frame->additional(), frame->additional_length());
-    block_addid_elem_size = EbmlElementSize(kMkvBlockAddID, frame->add_id());
+    block_additional_elem_size =
+        EbmlElementSize(libwebm::kMkvBlockAdditional, frame->additional(),
+                        frame->additional_length());
+    block_addid_elem_size =
+        EbmlElementSize(libwebm::kMkvBlockAddID, frame->add_id());

    block_more_payload_size =
        block_addid_elem_size + block_additional_elem_size;
    block_more_elem_size =
-        EbmlMasterElementSize(kMkvBlockMore, block_more_payload_size) +
+        EbmlMasterElementSize(libwebm::kMkvBlockMore, block_more_payload_size) +
        block_more_payload_size;
    block_additions_payload_size = block_more_elem_size;
    block_additions_elem_size =
-        EbmlMasterElementSize(kMkvBlockAdditions,
+        EbmlMasterElementSize(libwebm::kMkvBlockAdditions,
                              block_additions_payload_size) +
        block_additions_payload_size;
  }

-  uint64 discard_padding_elem_size = 0;
+  uint64_t discard_padding_elem_size = 0;
  if (frame->discard_padding() != 0) {
    discard_padding_elem_size =
-        EbmlElementSize(kMkvDiscardPadding, frame->discard_padding());
+        EbmlElementSize(libwebm::kMkvDiscardPadding, frame->discard_padding());
  }

-  const uint64 reference_block_timestamp =
+  const uint64_t reference_block_timestamp =
      frame->reference_block_timestamp() / timecode_scale;
-  uint64 reference_block_elem_size = 0;
+  uint64_t reference_block_elem_size = 0;
  if (!frame->is_key()) {
    reference_block_elem_size =
-        EbmlElementSize(kMkvReferenceBlock, reference_block_timestamp);
+        EbmlElementSize(libwebm::kMkvReferenceBlock, reference_block_timestamp);
  }

-  const uint64 duration = frame->duration() / timecode_scale;
-  uint64 block_duration_elem_size = 0;
+  const uint64_t duration = frame->duration() / timecode_scale;
+  uint64_t block_duration_elem_size = 0;
  if (duration > 0)
-    block_duration_elem_size = EbmlElementSize(kMkvBlockDuration, duration);
+    block_duration_elem_size =
+        EbmlElementSize(libwebm::kMkvBlockDuration, duration);

-  const uint64 block_payload_size = 4 + frame->length();
-  const uint64 block_elem_size =
-      EbmlMasterElementSize(kMkvBlock, block_payload_size) + block_payload_size;
+  const uint64_t block_payload_size = 4 + frame->length();
+  const uint64_t block_elem_size =
+      EbmlMasterElementSize(libwebm::kMkvBlock, block_payload_size) +
+      block_payload_size;

-  const uint64 block_group_payload_size =
+  const uint64_t block_group_payload_size =
      block_elem_size + block_additions_elem_size + block_duration_elem_size +
      discard_padding_elem_size + reference_block_elem_size;

-  if (!WriteEbmlMasterElement(writer, kMkvBlockGroup,
+  if (!WriteEbmlMasterElement(writer, libwebm::kMkvBlockGroup,
                              block_group_payload_size)) {
    return 0;
  }

-  if (!WriteEbmlMasterElement(writer, kMkvBlock, block_payload_size))
+  if (!WriteEbmlMasterElement(writer, libwebm::kMkvBlock, block_payload_size))
    return 0;

  if (WriteUInt(writer, frame->track_number()))
@@ -105,77 +105,81 @@ uint64 WriteBlock(IMkvWriter* writer, const Frame* const frame, int64 timecode,
  if (SerializeInt(writer, 0, 1))
    return 0;

-  if (writer->Write(frame->frame(), static_cast<uint32>(frame->length())))
+  if (writer->Write(frame->frame(), static_cast<uint32_t>(frame->length())))
    return 0;

  if (frame->additional()) {
-    if (!WriteEbmlMasterElement(writer, kMkvBlockAdditions,
+    if (!WriteEbmlMasterElement(writer, libwebm::kMkvBlockAdditions,
                                block_additions_payload_size)) {
      return 0;
    }

-    if (!WriteEbmlMasterElement(writer, kMkvBlockMore, block_more_payload_size))
+    if (!WriteEbmlMasterElement(writer, libwebm::kMkvBlockMore,
+                                block_more_payload_size))
      return 0;

-    if (!WriteEbmlElement(writer, kMkvBlockAddID, frame->add_id()))
+    if (!WriteEbmlElement(writer, libwebm::kMkvBlockAddID, frame->add_id()))
      return 0;

-    if (!WriteEbmlElement(writer, kMkvBlockAdditional, frame->additional(),
-                          frame->additional_length())) {
+    if (!WriteEbmlElement(writer, libwebm::kMkvBlockAdditional,
+                          frame->additional(), frame->additional_length())) {
      return 0;
    }
  }

  if (frame->discard_padding() != 0 &&
-      !WriteEbmlElement(writer, kMkvDiscardPadding, frame->discard_padding())) {
+      !WriteEbmlElement(writer, libwebm::kMkvDiscardPadding,
+                        frame->discard_padding())) {
    return false;
  }

  if (!frame->is_key() &&
-      !WriteEbmlElement(writer, kMkvReferenceBlock,
+      !WriteEbmlElement(writer, libwebm::kMkvReferenceBlock,
                        reference_block_timestamp)) {
    return false;
  }

-  if (duration > 0 && !WriteEbmlElement(writer, kMkvBlockDuration, duration)) {
+  if (duration > 0 &&
+      !WriteEbmlElement(writer, libwebm::kMkvBlockDuration, duration)) {
    return false;
  }
-  return EbmlMasterElementSize(kMkvBlockGroup, block_group_payload_size) +
+  return EbmlMasterElementSize(libwebm::kMkvBlockGroup,
+                               block_group_payload_size) +
         block_group_payload_size;
 }

-uint64 WriteSimpleBlock(IMkvWriter* writer, const Frame* const frame,
-                        int64 timecode) {
-  if (WriteID(writer, kMkvSimpleBlock))
+uint64_t WriteSimpleBlock(IMkvWriter* writer, const Frame* const frame,
+                          int64_t timecode) {
+  if (WriteID(writer, libwebm::kMkvSimpleBlock))
    return 0;

-  const int32 size = static_cast<int32>(frame->length()) + 4;
+  const int32_t size = static_cast<int32_t>(frame->length()) + 4;
  if (WriteUInt(writer, size))
    return 0;

-  if (WriteUInt(writer, static_cast<uint64>(frame->track_number())))
+  if (WriteUInt(writer, static_cast<uint64_t>(frame->track_number())))
    return 0;

  if (SerializeInt(writer, timecode, 2))
    return 0;

-  uint64 flags = 0;
+  uint64_t flags = 0;
  if (frame->is_key())
    flags |= 0x80;

  if (SerializeInt(writer, flags, 1))
    return 0;

-  if (writer->Write(frame->frame(), static_cast<uint32>(frame->length())))
+  if (writer->Write(frame->frame(), static_cast<uint32_t>(frame->length())))
    return 0;

-  return GetUIntSize(kMkvSimpleBlock) + GetCodedUIntSize(size) + 4 +
-         frame->length();
+  return static_cast<uint64_t>(GetUIntSize(libwebm::kMkvSimpleBlock) +
+                               GetCodedUIntSize(size) + 4 + frame->length());
 }

 }  // namespace

-int32 GetCodedUIntSize(uint64 value) {
+int32_t GetCodedUIntSize(uint64_t value) {
  if (value < 0x000000000000007FULL)
    return 1;
  else if (value < 0x0000000000003FFFULL)
@@ -193,7 +197,7 @@ int32 GetCodedUIntSize(uint64 value) {
  return 8;
 }

-int32 GetUIntSize(uint64 value) {
+int32_t GetUIntSize(uint64_t value) {
  if (value < 0x0000000000000100ULL)
    return 1;
  else if (value < 0x0000000000010000ULL)
@@ -211,26 +215,26 @@ int32 GetUIntSize(uint64 value) {
  return 8;
 }

-int32 GetIntSize(int64 value) {
+int32_t GetIntSize(int64_t value) {
  // Doubling the requested value ensures positive values with their high bit
  // set are written with 0-padding to avoid flipping the signedness.
-  const uint64 v = (value < 0) ? value ^ -1LL : value;
+  const uint64_t v = (value < 0) ? value ^ -1LL : value;
  return GetUIntSize(2 * v);
 }

-uint64 EbmlMasterElementSize(uint64 type, uint64 value) {
+uint64_t EbmlMasterElementSize(uint64_t type, uint64_t value) {
  // Size of EBML ID
-  int32 ebml_size = GetUIntSize(type);
+  int32_t ebml_size = GetUIntSize(type);

  // Datasize
  ebml_size += GetCodedUIntSize(value);

-  return ebml_size;
+  return static_cast<uint64_t>(ebml_size);
 }

-uint64 EbmlElementSize(uint64 type, int64 value) {
+uint64_t EbmlElementSize(uint64_t type, int64_t value) {
  // Size of EBML ID
-  int32 ebml_size = GetUIntSize(type);
+  int32_t ebml_size = GetUIntSize(type);

  // Datasize
  ebml_size += GetIntSize(value);
@@ -238,15 +242,20 @@ uint64 EbmlElementSize(uint64 type, int64 value) {
  // Size of Datasize
  ebml_size++;

-  return ebml_size;
+  return static_cast<uint64_t>(ebml_size);
 }

-uint64 EbmlElementSize(uint64 type, uint64 value) {
+uint64_t EbmlElementSize(uint64_t type, uint64_t value) {
+  return EbmlElementSize(type, value, 0);
+}
+
+uint64_t EbmlElementSize(uint64_t type, uint64_t value, uint64_t fixed_size) {
  // Size of EBML ID
-  int32 ebml_size = GetUIntSize(type);
+  uint64_t ebml_size = static_cast<uint64_t>(GetUIntSize(type));

  // Datasize
-  ebml_size += GetUIntSize(value);
+  ebml_size +=
+      (fixed_size > 0) ? fixed_size : static_cast<uint64_t>(GetUIntSize(value));

  // Size of Datasize
  ebml_size++;
@@ -254,9 +263,9 @@ uint64 EbmlElementSize(uint64 type, uint64 value) {
  return ebml_size;
 }

-uint64 EbmlElementSize(uint64 type, float /* value */) {
+uint64_t EbmlElementSize(uint64_t type, float /* value */) {
  // Size of EBML ID
-  uint64 ebml_size = GetUIntSize(type);
+  uint64_t ebml_size = static_cast<uint64_t>(GetUIntSize(type));

  // Datasize
  ebml_size += sizeof(float);
@@ -267,12 +276,12 @@ uint64 EbmlElementSize(uint64 type, float /* value */) {
  return ebml_size;
 }

-uint64 EbmlElementSize(uint64 type, const char* value) {
+uint64_t EbmlElementSize(uint64_t type, const char* value) {
  if (!value)
    return 0;

  // Size of EBML ID
-  uint64 ebml_size = GetUIntSize(type);
+  uint64_t ebml_size = static_cast<uint64_t>(GetUIntSize(type));

  // Datasize
  ebml_size += strlen(value);
@@ -283,12 +292,12 @@ uint64 EbmlElementSize(uint64 type, const char* value) {
  return ebml_size;
 }

-uint64 EbmlElementSize(uint64 type, const uint8* value, uint64 size) {
+uint64_t EbmlElementSize(uint64_t type, const uint8_t* value, uint64_t size) {
  if (!value)
    return 0;

  // Size of EBML ID
-  uint64 ebml_size = GetUIntSize(type);
+  uint64_t ebml_size = static_cast<uint64_t>(GetUIntSize(type));

  // Datasize
  ebml_size += size;
@@ -299,9 +308,9 @@ uint64 EbmlElementSize(uint64 type, const uint8* value, uint64 size) {
  return ebml_size;
 }

-uint64 EbmlDateElementSize(uint64 type) {
+uint64_t EbmlDateElementSize(uint64_t type) {
  // Size of EBML ID
-  uint64 ebml_size = GetUIntSize(type);
+  uint64_t ebml_size = static_cast<uint64_t>(GetUIntSize(type));

  // Datasize
  ebml_size += kDateElementSize;
@@ -312,18 +321,18 @@ uint64 EbmlDateElementSize(uint64 type) {
  return ebml_size;
 }

-int32 SerializeInt(IMkvWriter* writer, int64 value, int32 size) {
+int32_t SerializeInt(IMkvWriter* writer, int64_t value, int32_t size) {
  if (!writer || size < 1 || size > 8)
    return -1;

-  for (int32 i = 1; i <= size; ++i) {
-    const int32 byte_count = size - i;
-    const int32 bit_count = byte_count * 8;
+  for (int32_t i = 1; i <= size; ++i) {
+    const int32_t byte_count = size - i;
+    const int32_t bit_count = byte_count * 8;

-    const int64 bb = value >> bit_count;
-    const uint8 b = static_cast<uint8>(bb);
+    const int64_t bb = value >> bit_count;
+    const uint8_t b = static_cast<uint8_t>(bb);

-    const int32 status = writer->Write(&b, 1);
+    const int32_t status = writer->Write(&b, 1);

    if (status < 0)
      return status;
@@ -332,26 +341,26 @@ int32 SerializeInt(IMkvWriter* writer, int64 value, int32 size) {
  return 0;
 }

-int32 SerializeFloat(IMkvWriter* writer, float f) {
+int32_t SerializeFloat(IMkvWriter* writer, float f) {
  if (!writer)
    return -1;

-  assert(sizeof(uint32) == sizeof(float));
+  assert(sizeof(uint32_t) == sizeof(float));
  // This union is merely used to avoid a reinterpret_cast from float& to
  // uint32& which will result in violation of strict aliasing.
  union U32 {
-    uint32 u32;
+    uint32_t u32;
    float f;
  } value;
  value.f = f;

-  for (int32 i = 1; i <= 4; ++i) {
-    const int32 byte_count = 4 - i;
-    const int32 bit_count = byte_count * 8;
+  for (int32_t i = 1; i <= 4; ++i) {
+    const int32_t byte_count = 4 - i;
+    const int32_t bit_count = byte_count * 8;

-    const uint8 byte = static_cast<uint8>(value.u32 >> bit_count);
+    const uint8_t byte = static_cast<uint8_t>(value.u32 >> bit_count);

-    const int32 status = writer->Write(&byte, 1);
+    const int32_t status = writer->Write(&byte, 1);

    if (status < 0)
      return status;
@@ -360,21 +369,21 @@ int32 SerializeFloat(IMkvWriter* writer, float f) {
  return 0;
 }

-int32 WriteUInt(IMkvWriter* writer, uint64 value) {
+int32_t WriteUInt(IMkvWriter* writer, uint64_t value) {
  if (!writer)
    return -1;

-  int32 size = GetCodedUIntSize(value);
+  int32_t size = GetCodedUIntSize(value);

  return WriteUIntSize(writer, value, size);
 }

-int32 WriteUIntSize(IMkvWriter* writer, uint64 value, int32 size) {
+int32_t WriteUIntSize(IMkvWriter* writer, uint64_t value, int32_t size) {
  if (!writer || size < 0 || size > 8)
    return -1;

  if (size > 0) {
-    const uint64 bit = 1LL << (size * 7);
+    const uint64_t bit = 1LL << (size * 7);

    if (value > (bit - 2))
      return -1;
@@ -382,11 +391,11 @@ int32 WriteUIntSize(IMkvWriter* writer, uint64 value, int32 size) {
    value |= bit;
  } else {
    size = 1;
-    int64 bit;
+    int64_t bit;

    for (;;) {
      bit = 1LL << (size * 7);
-      const uint64 max = bit - 2;
+      const uint64_t max = bit - 2;

      if (value <= max)
        break;
@@ -403,18 +412,18 @@ int32 WriteUIntSize(IMkvWriter* writer, uint64 value, int32 size) {
  return SerializeInt(writer, value, size);
 }

-int32 WriteID(IMkvWriter* writer, uint64 type) {
+int32_t WriteID(IMkvWriter* writer, uint64_t type) {
  if (!writer)
    return -1;

  writer->ElementStartNotify(type, writer->Position());

-  const int32 size = GetUIntSize(type);
+  const int32_t size = GetUIntSize(type);

  return SerializeInt(writer, type, size);
 }

-bool WriteEbmlMasterElement(IMkvWriter* writer, uint64 type, uint64 size) {
+bool WriteEbmlMasterElement(IMkvWriter* writer, uint64_t type, uint64_t size) {
  if (!writer)
    return false;

@@ -427,41 +436,51 @@ bool WriteEbmlMasterElement(IMkvWriter* writer, uint64 type, uint64 size) {
  return true;
 }

-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, uint64 value) {
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, uint64_t value) {
+  return WriteEbmlElement(writer, type, value, 0);
+}
+
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, uint64_t value,
+                      uint64_t fixed_size) {
  if (!writer)
    return false;

  if (WriteID(writer, type))
    return false;

-  const uint64 size = GetUIntSize(value);
+  uint64_t size = static_cast<uint64_t>(GetUIntSize(value));
+  if (fixed_size > 0) {
+    if (size > fixed_size)
+      return false;
+    size = fixed_size;
+  }
  if (WriteUInt(writer, size))
    return false;

-  if (SerializeInt(writer, value, static_cast<int32>(size)))
+  if (SerializeInt(writer, value, static_cast<int32_t>(size)))
    return false;

  return true;
 }

-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, int64 value) {
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, int64_t value) {
  if (!writer)
    return false;

  if (WriteID(writer, type))
    return 0;

-  const uint64 size = GetIntSize(value);
+  const uint64_t size = GetIntSize(value);
  if (WriteUInt(writer, size))
    return false;

-  if (SerializeInt(writer, value, static_cast<int32>(size)))
+  if (SerializeInt(writer, value, static_cast<int32_t>(size)))
    return false;

  return true;
 }

-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, float value) {
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, float value) {
  if (!writer)
    return false;

@@ -477,25 +496,25 @@ bool WriteEbmlElement(IMkvWriter* writer, uint64 type, float value) {
  return true;
 }

-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, const char* value) {
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, const char* value) {
  if (!writer || !value)
    return false;

  if (WriteID(writer, type))
    return false;

-  const uint64 length = strlen(value);
+  const uint64_t length = strlen(value);
  if (WriteUInt(writer, length))
    return false;

-  if (writer->Write(value, static_cast<const uint32>(length)))
+  if (writer->Write(value, static_cast<const uint32_t>(length)))
    return false;

  return true;
 }

-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, const uint8* value,
-                      uint64 size) {
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, const uint8_t* value,
+                      uint64_t size) {
  if (!writer || !value || size < 1)
    return false;

@@ -505,13 +524,13 @@ bool WriteEbmlElement(IMkvWriter* writer, uint64 type, const uint8* value,
  if (WriteUInt(writer, size))
    return false;

-  if (writer->Write(value, static_cast<uint32>(size)))
+  if (writer->Write(value, static_cast<uint32_t>(size)))
    return false;

  return true;
 }

-bool WriteEbmlDateElement(IMkvWriter* writer, uint64 type, int64 value) {
+bool WriteEbmlDateElement(IMkvWriter* writer, uint64_t type, int64_t value) {
  if (!writer)
    return false;

@@ -527,8 +546,8 @@ bool WriteEbmlDateElement(IMkvWriter* writer, uint64 type, int64 value) {
  return true;
 }

-uint64 WriteFrame(IMkvWriter* writer, const Frame* const frame,
-                  Cluster* cluster) {
+uint64_t WriteFrame(IMkvWriter* writer, const Frame* const frame,
+                    Cluster* cluster) {
  if (!writer || !frame || !frame->IsValid() || !cluster ||
      !cluster->timecode_scale())
    return 0;
@@ -537,7 +556,7 @@ uint64 WriteFrame(IMkvWriter* writer, const Frame* const frame,
  //  timecode for the cluster itself (remember that block timecode
  //  is a signed, 16-bit integer).  However, as a simplification we
  //  only permit non-negative cluster-relative timecodes for blocks.
-  const int64 relative_timecode = cluster->GetRelativeTimecode(
+  const int64_t relative_timecode = cluster->GetRelativeTimecode(
      frame->timestamp() / cluster->timecode_scale());
  if (relative_timecode < 0 || relative_timecode > kMaxBlockTimecode)
    return 0;
@@ -548,53 +567,53 @@ uint64 WriteFrame(IMkvWriter* writer, const Frame* const frame,
                        cluster->timecode_scale());
 }

-uint64 WriteVoidElement(IMkvWriter* writer, uint64 size) {
+uint64_t WriteVoidElement(IMkvWriter* writer, uint64_t size) {
  if (!writer)
    return false;

  // Subtract one for the void ID and the coded size.
-  uint64 void_entry_size = size - 1 - GetCodedUIntSize(size - 1);
-  uint64 void_size =
-      EbmlMasterElementSize(kMkvVoid, void_entry_size) + void_entry_size;
+  uint64_t void_entry_size = size - 1 - GetCodedUIntSize(size - 1);
+  uint64_t void_size =
+      EbmlMasterElementSize(libwebm::kMkvVoid, void_entry_size) +
+      void_entry_size;

  if (void_size != size)
    return 0;

-  const int64 payload_position = writer->Position();
+  const int64_t payload_position = writer->Position();
  if (payload_position < 0)
    return 0;

-  if (WriteID(writer, kMkvVoid))
+  if (WriteID(writer, libwebm::kMkvVoid))
    return 0;

  if (WriteUInt(writer, void_entry_size))
    return 0;

-  const uint8 value = 0;
-  for (int32 i = 0; i < static_cast<int32>(void_entry_size); ++i) {
+  const uint8_t value = 0;
+  for (int32_t i = 0; i < static_cast<int32_t>(void_entry_size); ++i) {
    if (writer->Write(&value, 1))
      return 0;
  }

-  const int64 stop_position = writer->Position();
+  const int64_t stop_position = writer->Position();
  if (stop_position < 0 ||
-      stop_position - payload_position != static_cast<int64>(void_size))
+      stop_position - payload_position != static_cast<int64_t>(void_size))
    return 0;

  return void_size;
 }

-void GetVersion(int32* major, int32* minor, int32* build, int32* revision) {
+void GetVersion(int32_t* major, int32_t* minor, int32_t* build,
+                int32_t* revision) {
  *major = 0;
  *minor = 2;
  *build = 1;
  *revision = 0;
 }

-}  // namespace mkvmuxer
-
-mkvmuxer::uint64 mkvmuxer::MakeUID(unsigned int* seed) {
-  uint64 uid = 0;
+uint64_t MakeUID(unsigned int* seed) {
+  uint64_t uid = 0;

 #ifdef __MINGW32__
  srand(*seed);
@@ -606,24 +625,26 @@ mkvmuxer::uint64 mkvmuxer::MakeUID(unsigned int* seed) {
 // TODO(fgalligan): Move random number generation to platform specific code.
 #ifdef _MSC_VER
    (void)seed;
-    const int32 nn = rand();
+    const int32_t nn = rand();
 #elif __ANDROID__
-    int32 temp_num = 1;
+    int32_t temp_num = 1;
    int fd = open("/dev/urandom", O_RDONLY);
    if (fd != -1) {
-      read(fd, &temp_num, sizeof(int32));
+      read(fd, &temp_num, sizeof(temp_num));
      close(fd);
    }
-    const int32 nn = temp_num;
+    const int32_t nn = temp_num;
 #elif defined __MINGW32__
-    const int32 nn = rand();
+    const int32_t nn = rand();
 #else
-    const int32 nn = rand_r(seed);
+    const int32_t nn = rand_r(seed);
 #endif
-    const int32 n = 0xFF & (nn >> 4);  // throw away low-order bits
+    const int32_t n = 0xFF & (nn >> 4);  // throw away low-order bits

    uid |= n;
  }

  return uid;
 }
+
+}  // namespace mkvmuxer
--- a/third_party/libwebm/mkvmuxer/mkvmuxerutil.h
+++ b/third_party/libwebm/mkvmuxer/mkvmuxerutil.h
@@ -0,0 +1,95 @@
+// Copyright (c) 2012 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 MKVMUXER_MKVMUXERUTIL_H_
+#define MKVMUXER_MKVMUXERUTIL_H_
+
+#include <stdint.h>
+
+namespace mkvmuxer {
+class Cluster;
+class Frame;
+class IMkvWriter;
+
+const uint64_t kEbmlUnknownValue = 0x01FFFFFFFFFFFFFFULL;
+const int64_t kMaxBlockTimecode = 0x07FFFLL;
+
+// Writes out |value| in Big Endian order. Returns 0 on success.
+int32_t SerializeInt(IMkvWriter* writer, int64_t value, int32_t size);
+
+// Returns the size in bytes of the element.
+int32_t GetUIntSize(uint64_t value);
+int32_t GetIntSize(int64_t value);
+int32_t GetCodedUIntSize(uint64_t value);
+uint64_t EbmlMasterElementSize(uint64_t type, uint64_t value);
+uint64_t EbmlElementSize(uint64_t type, int64_t value);
+uint64_t EbmlElementSize(uint64_t type, uint64_t value);
+uint64_t EbmlElementSize(uint64_t type, float value);
+uint64_t EbmlElementSize(uint64_t type, const char* value);
+uint64_t EbmlElementSize(uint64_t type, const uint8_t* value, uint64_t size);
+uint64_t EbmlDateElementSize(uint64_t type);
+
+// Returns the size in bytes of the element assuming that the element was
+// written using |fixed_size| bytes. If |fixed_size| is set to zero, then it
+// computes the necessary number of bytes based on |value|.
+uint64_t EbmlElementSize(uint64_t type, uint64_t value, uint64_t fixed_size);
+
+// Creates an EBML coded number from |value| and writes it out. The size of
+// the coded number is determined by the value of |value|. |value| must not
+// be in a coded form. Returns 0 on success.
+int32_t WriteUInt(IMkvWriter* writer, uint64_t value);
+
+// Creates an EBML coded number from |value| and writes it out. The size of
+// the coded number is determined by the value of |size|. |value| must not
+// be in a coded form. Returns 0 on success.
+int32_t WriteUIntSize(IMkvWriter* writer, uint64_t value, int32_t size);
+
+// Output an Mkv master element. Returns true if the element was written.
+bool WriteEbmlMasterElement(IMkvWriter* writer, uint64_t value, uint64_t size);
+
+// Outputs an Mkv ID, calls |IMkvWriter::ElementStartNotify|, and passes the
+// ID to |SerializeInt|. Returns 0 on success.
+int32_t WriteID(IMkvWriter* writer, uint64_t type);
+
+// Output an Mkv non-master element. Returns true if the element was written.
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, uint64_t value);
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, int64_t value);
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, float value);
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, const char* value);
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, const uint8_t* value,
+                      uint64_t size);
+bool WriteEbmlDateElement(IMkvWriter* writer, uint64_t type, int64_t value);
+
+// Output an Mkv non-master element using fixed size. The element will be
+// written out using exactly |fixed_size| bytes. If |fixed_size| is set to zero
+// then it computes the necessary number of bytes based on |value|. Returns true
+// if the element was written.
+bool WriteEbmlElement(IMkvWriter* writer, uint64_t type, uint64_t value,
+                      uint64_t fixed_size);
+
+// Output a Mkv Frame. It decides the correct element to write (Block vs
+// SimpleBlock) based on the parameters of the Frame.
+uint64_t WriteFrame(IMkvWriter* writer, const Frame* const frame,
+                    Cluster* cluster);
+
+// Output a void element. |size| must be the entire size in bytes that will be
+// void. The function will calculate the size of the void header and subtract
+// it from |size|.
+uint64_t WriteVoidElement(IMkvWriter* writer, uint64_t size);
+
+// Returns the version number of the muxer in |major|, |minor|, |build|,
+// and |revision|.
+void GetVersion(int32_t* major, int32_t* minor, int32_t* build,
+                int32_t* revision);
+
+// Returns a random number to be used for UID, using |seed| to seed
+// the random-number generator (see POSIX rand_r() for semantics).
+uint64_t MakeUID(unsigned int* seed);
+
+}  // namespace mkvmuxer
+
+#endif  // MKVMUXER_MKVMUXERUTIL_H_
--- a/third_party/libwebm/mkvmuxer/mkvwriter.cc
+++ b/third_party/libwebm/mkvmuxer/mkvwriter.cc
@@ -6,14 +6,12 @@
 // in the file PATENTS.  All contributing project authors may
 // be found in the AUTHORS file in the root of the source tree.

-#include "mkvwriter.hpp"
+#include "mkvmuxer/mkvwriter.h"

 #ifdef _MSC_VER
 #include <share.h>  // for _SH_DENYWR
 #endif

-#include <new>
-
 namespace mkvmuxer {

 MkvWriter::MkvWriter() : file_(NULL), writer_owns_file_(true) {}
--- a/third_party/libwebm/mkvmuxer/mkvwriter.h
+++ b/third_party/libwebm/mkvmuxer/mkvwriter.h
@@ -6,13 +6,13 @@
 // in the file PATENTS.  All contributing project authors may
 // be found in the AUTHORS file in the root of the source tree.

-#ifndef MKVWRITER_HPP
-#define MKVWRITER_HPP
+#ifndef MKVMUXER_MKVWRITER_H_
+#define MKVMUXER_MKVWRITER_H_

 #include <stdio.h>

-#include "mkvmuxer.hpp"
-#include "mkvmuxertypes.hpp"
+#include "mkvmuxer/mkvmuxer.h"
+#include "mkvmuxer/mkvmuxertypes.h"

 namespace mkvmuxer {

@@ -46,6 +46,6 @@ class MkvWriter : public IMkvWriter {
  LIBWEBM_DISALLOW_COPY_AND_ASSIGN(MkvWriter);
 };

-}  // end namespace mkvmuxer
+}  // namespace mkvmuxer

-#endif  // MKVWRITER_HPP
+#endif  // MKVMUXER_MKVWRITER_H_
--- a/third_party/libwebm/mkvmuxerutil.hpp
+++ b/third_party/libwebm/mkvmuxerutil.hpp
@@ -1,83 +0,0 @@
-// Copyright (c) 2012 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 MKVMUXERUTIL_HPP
-#define MKVMUXERUTIL_HPP
-
-#include "mkvmuxer.hpp"
-#include "mkvmuxertypes.hpp"
-
-namespace mkvmuxer {
-
-class IMkvWriter;
-
-const uint64 kEbmlUnknownValue = 0x01FFFFFFFFFFFFFFULL;
-const int64 kMaxBlockTimecode = 0x07FFFLL;
-
-// Writes out |value| in Big Endian order. Returns 0 on success.
-int32 SerializeInt(IMkvWriter* writer, int64 value, int32 size);
-
-// Returns the size in bytes of the element.
-int32 GetUIntSize(uint64 value);
-int32 GetIntSize(int64 value);
-int32 GetCodedUIntSize(uint64 value);
-uint64 EbmlMasterElementSize(uint64 type, uint64 value);
-uint64 EbmlElementSize(uint64 type, int64 value);
-uint64 EbmlElementSize(uint64 type, uint64 value);
-uint64 EbmlElementSize(uint64 type, float value);
-uint64 EbmlElementSize(uint64 type, const char* value);
-uint64 EbmlElementSize(uint64 type, const uint8* value, uint64 size);
-uint64 EbmlDateElementSize(uint64 type);
-
-// Creates an EBML coded number from |value| and writes it out. The size of
-// the coded number is determined by the value of |value|. |value| must not
-// be in a coded form. Returns 0 on success.
-int32 WriteUInt(IMkvWriter* writer, uint64 value);
-
-// Creates an EBML coded number from |value| and writes it out. The size of
-// the coded number is determined by the value of |size|. |value| must not
-// be in a coded form. Returns 0 on success.
-int32 WriteUIntSize(IMkvWriter* writer, uint64 value, int32 size);
-
-// Output an Mkv master element. Returns true if the element was written.
-bool WriteEbmlMasterElement(IMkvWriter* writer, uint64 value, uint64 size);
-
-// Outputs an Mkv ID, calls |IMkvWriter::ElementStartNotify|, and passes the
-// ID to |SerializeInt|. Returns 0 on success.
-int32 WriteID(IMkvWriter* writer, uint64 type);
-
-// Output an Mkv non-master element. Returns true if the element was written.
-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, uint64 value);
-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, int64 value);
-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, float value);
-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, const char* value);
-bool WriteEbmlElement(IMkvWriter* writer, uint64 type, const uint8* value,
-                      uint64 size);
-bool WriteEbmlDateElement(IMkvWriter* writer, uint64 type, int64 value);
-
-// Output a Mkv Frame. It decides the correct element to write (Block vs
-// SimpleBlock) based on the parameters of the Frame.
-uint64 WriteFrame(IMkvWriter* writer, const Frame* const frame,
-                  Cluster* cluster);
-
-// Output a void element. |size| must be the entire size in bytes that will be
-// void. The function will calculate the size of the void header and subtract
-// it from |size|.
-uint64 WriteVoidElement(IMkvWriter* writer, uint64 size);
-
-// Returns the version number of the muxer in |major|, |minor|, |build|,
-// and |revision|.
-void GetVersion(int32* major, int32* minor, int32* build, int32* revision);
-
-// Returns a random number to be used for UID, using |seed| to seed
-// the random-number generator (see POSIX rand_r() for semantics).
-uint64 MakeUID(unsigned int* seed);
-
-}  // end namespace mkvmuxer
-
-#endif  // MKVMUXERUTIL_HPP
--- a/third_party/libwebm/mkvparser/mkvparser.cc
+++ b/third_party/libwebm/mkvparser/mkvparser.cc
--- a/third_party/libwebm/mkvparser/mkvparser.h
+++ b/third_party/libwebm/mkvparser/mkvparser.h
@@ -5,13 +5,10 @@
 // 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 MKVPARSER_HPP
-#define MKVPARSER_HPP
+#ifndef MKVPARSER_MKVPARSER_H_
+#define MKVPARSER_MKVPARSER_H_

 #include <cstddef>
-#include <cstdio>
-#include <cstdlib>

 namespace mkvparser {

@@ -28,8 +25,9 @@ class IMkvReader {
  virtual ~IMkvReader();
 };

-template<typename Type> Type* SafeArrayAlloc(unsigned long long num_elements,
-                                             unsigned long long element_size);
+template <typename Type>
+Type* SafeArrayAlloc(unsigned long long num_elements,
+                     unsigned long long element_size);
 long long GetUIntLength(IMkvReader*, long long, long&);
 long long ReadUInt(IMkvReader*, long long, long&);
 long long ReadID(IMkvReader* pReader, long long pos, long& len);
@@ -128,7 +126,7 @@ class BlockEntry {
 public:
  virtual ~BlockEntry();

-  bool EOS() const;
+  bool EOS() const { return (GetKind() == kBlockEOS); }
  const Cluster* GetCluster() const;
  long GetIndex() const;
  virtual const Block* GetBlock() const = 0;
@@ -391,6 +389,90 @@ class Track {
  ContentEncoding** content_encoding_entries_end_;
 };

+struct PrimaryChromaticity {
+  PrimaryChromaticity() : x(0), y(0) {}
+  ~PrimaryChromaticity() {}
+  static bool Parse(IMkvReader* reader, long long read_pos,
+                    long long value_size, bool is_x,
+                    PrimaryChromaticity** chromaticity);
+  float x;
+  float y;
+};
+
+struct MasteringMetadata {
+  static const float kValueNotPresent;
+
+  MasteringMetadata()
+      : r(NULL),
+        g(NULL),
+        b(NULL),
+        white_point(NULL),
+        luminance_max(kValueNotPresent),
+        luminance_min(kValueNotPresent) {}
+  ~MasteringMetadata() {
+    delete r;
+    delete g;
+    delete b;
+    delete white_point;
+  }
+
+  static bool Parse(IMkvReader* reader, long long element_start,
+                    long long element_size,
+                    MasteringMetadata** mastering_metadata);
+
+  PrimaryChromaticity* r;
+  PrimaryChromaticity* g;
+  PrimaryChromaticity* b;
+  PrimaryChromaticity* white_point;
+  float luminance_max;
+  float luminance_min;
+};
+
+struct Colour {
+  static const long long kValueNotPresent;
+
+  // Unless otherwise noted all values assigned upon construction are the
+  // equivalent of unspecified/default.
+  Colour()
+      : matrix_coefficients(kValueNotPresent),
+        bits_per_channel(kValueNotPresent),
+        chroma_subsampling_horz(kValueNotPresent),
+        chroma_subsampling_vert(kValueNotPresent),
+        cb_subsampling_horz(kValueNotPresent),
+        cb_subsampling_vert(kValueNotPresent),
+        chroma_siting_horz(kValueNotPresent),
+        chroma_siting_vert(kValueNotPresent),
+        range(kValueNotPresent),
+        transfer_characteristics(kValueNotPresent),
+        primaries(kValueNotPresent),
+        max_cll(kValueNotPresent),
+        max_fall(kValueNotPresent),
+        mastering_metadata(NULL) {}
+  ~Colour() {
+    delete mastering_metadata;
+    mastering_metadata = NULL;
+  }
+
+  static bool Parse(IMkvReader* reader, long long element_start,
+                    long long element_size, Colour** colour);
+
+  long long matrix_coefficients;
+  long long bits_per_channel;
+  long long chroma_subsampling_horz;
+  long long chroma_subsampling_vert;
+  long long cb_subsampling_horz;
+  long long cb_subsampling_vert;
+  long long chroma_siting_horz;
+  long long chroma_siting_vert;
+  long long range;
+  long long transfer_characteristics;
+  long long primaries;
+  long long max_cll;
+  long long max_fall;
+
+  MasteringMetadata* mastering_metadata;
+};
+
 class VideoTrack : public Track {
  VideoTrack(const VideoTrack&);
  VideoTrack& operator=(const VideoTrack&);
@@ -398,6 +480,7 @@ class VideoTrack : public Track {
  VideoTrack(Segment*, long long element_start, long long element_size);

 public:
+  virtual ~VideoTrack();
  static long Parse(Segment*, const Info&, long long element_start,
                    long long element_size, VideoTrack*&);

@@ -412,6 +495,8 @@ class VideoTrack : public Track {
  bool VetEntry(const BlockEntry*) const;
  long Seek(long long time_ns, const BlockEntry*&) const;

+  Colour* GetColour() const;
+
 private:
  long long m_width;
  long long m_height;
@@ -421,6 +506,8 @@ class VideoTrack : public Track {
  long long m_stereo_mode;

  double m_rate;
+
+  Colour* m_colour;
 };

 class AudioTrack : public Track {
@@ -1013,7 +1100,7 @@ class Segment {
  const BlockEntry* GetBlock(const CuePoint&, const CuePoint::TrackPosition&);
 };

-}  // end namespace mkvparser
+}  // namespace mkvparser

 inline long mkvparser::Segment::LoadCluster() {
  long long pos;
@@ -1022,4 +1109,4 @@ inline long mkvparser::Segment::LoadCluster() {
  return LoadCluster(pos, size);
 }

-#endif  // MKVPARSER_HPP
+#endif  // MKVPARSER_MKVPARSER_H_
--- a/third_party/libwebm/mkvparser/mkvreader.cc
+++ b/third_party/libwebm/mkvparser/mkvreader.cc
@@ -5,8 +5,7 @@
 // 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 "mkvreader.hpp"
+#include "mkvparser/mkvreader.h"

 #include <cassert>

@@ -129,4 +128,4 @@ int MkvReader::Read(long long offset, long len, unsigned char* buffer) {
  return 0;  // success
 }

-}  // end namespace mkvparser
+}  // namespace mkvparser
--- a/third_party/libwebm/mkvparser/mkvreader.h
+++ b/third_party/libwebm/mkvparser/mkvreader.h
@@ -5,13 +5,13 @@
 // 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 MKVPARSER_MKVREADER_H_
+#define MKVPARSER_MKVREADER_H_

-#ifndef MKVREADER_HPP
-#define MKVREADER_HPP
-
-#include "mkvparser.hpp"
 #include <cstdio>

+#include "mkvparser/mkvparser.h"
+
 namespace mkvparser {

 class MkvReader : public IMkvReader {
@@ -40,6 +40,6 @@ class MkvReader : public IMkvReader {
  bool reader_owns_file_;
 };

-}  // end namespace mkvparser
+}  // namespace mkvparser

-#endif  // MKVREADER_HPP
+#endif  // MKVPARSER_MKVREADER_H_
--- a/third_party/x86inc/README.libvpx
+++ b/third_party/x86inc/README.libvpx
@@ -1,5 +1,5 @@
-URL: http://git.videolan.org/?p=x264.git
-Version: a95584945dd9ce3acc66c6cd8f6796bc4404d40d
+URL: https://git.videolan.org/git/x264.git
+Version: d23d18655249944c1ca894b451e2c82c7a584c62
 License: ISC
 License File: LICENSE

@@ -13,12 +13,8 @@ Prefix functions with vpx by default.
 Manage name mangling (prefixing with '_') manually because 'PREFIX' does not
  exist in libvpx.
 Expand PIC default to macho64 and respect CONFIG_PIC from libvpx
-Catch all elf formats for 'hidden' status and SECTION notes.
-Avoid 'amdnop' when building with nasm.
 Set 'private_extern' visibility for macho targets.
 Copy PIC 'GLOBAL' macros from x86_abi_support.asm
 Use .text instead of .rodata on macho to avoid broken tables in PIC mode.
 Use .text with no alignment for aout
 Only use 'hidden' visibility with Chromium
-Move '%use smartalign' for nasm out of 'INIT_CPUFLAGS' and before
-  'ALIGNMODE'.
--- a/third_party/x86inc/x86inc.asm
+++ b/third_party/x86inc/x86inc.asm
@@ -1,7 +1,7 @@
 ;*****************************************************************************
 ;* x86inc.asm: x264asm abstraction layer
 ;*****************************************************************************
-;* Copyright (C) 2005-2015 x264 project
+;* Copyright (C) 2005-2016 x264 project
 ;*
 ;* Authors: Loren Merritt <lorenm@u.washington.edu>
 ;*          Anton Mitrofanov <BugMaster@narod.ru>
@@ -66,16 +66,35 @@
    %endif
 %endif

-%ifidn   __OUTPUT_FORMAT__,elf32
-    %define mangle(x) x
+%define FORMAT_ELF 0
+%ifidn __OUTPUT_FORMAT__,elf
+    %define FORMAT_ELF 1
+%elifidn __OUTPUT_FORMAT__,elf32
+    %define FORMAT_ELF 1
 %elifidn __OUTPUT_FORMAT__,elf64
-    %define mangle(x) x
-%elifidn __OUTPUT_FORMAT__,x64
-    %define mangle(x) x
-%elifidn __OUTPUT_FORMAT__,win64
-    %define mangle(x) x
+    %define FORMAT_ELF 1
+%endif
+
+%define FORMAT_MACHO 0
+%ifidn __OUTPUT_FORMAT__,macho32
+     %define FORMAT_MACHO 1
+%elifidn __OUTPUT_FORMAT__,macho64
+     %define FORMAT_MACHO 1
+%endif
+
+; Set PREFIX for libvpx builds.
+%if FORMAT_ELF
+    %undef PREFIX
+%elif WIN64
+    %undef PREFIX
 %else
+    %define PREFIX
+%endif
+
+%ifdef PREFIX
    %define mangle(x) _ %+ x
+%else
+    %define mangle(x) x
 %endif

 ; In some instances macho32 tables get misaligned when using .rodata.
@@ -94,14 +113,6 @@
    %endif
 %endmacro

-%macro SECTION_TEXT 0-1 16
-    %ifidn __OUTPUT_FORMAT__,aout
-        SECTION .text
-    %else
-        SECTION .text align=%1
-    %endif
-%endmacro
-
 ; PIC macros are copied from vpx_ports/x86_abi_support.asm. The "define PIC"
 ; from original code is added in for 64bit.
 %ifidn __OUTPUT_FORMAT__,elf32
@@ -188,8 +199,16 @@
 %ifdef PIC
    default rel
 %endif
+
+%ifndef GET_GOT_DEFINED
+    %define GET_GOT_DEFINED 0
+%endif
 ; Done with PIC macros

+%ifdef __NASM_VER__
+    %use smartalign
+%endif
+
 ; Macros to eliminate most code duplication between x86_32 and x86_64:
 ; Currently this works only for leaf functions which load all their arguments
 ; into registers at the start, and make no other use of the stack. Luckily that
@@ -237,6 +256,7 @@
    %define r%1w %2w
    %define r%1b %2b
    %define r%1h %2h
+    %define %2q %2
    %if %0 == 2
        %define r%1m  %2d
        %define r%1mp %2
@@ -261,9 +281,9 @@
    %define e%1h %3
    %define r%1b %2
    %define e%1b %2
-%if ARCH_X86_64 == 0
-    %define r%1  e%1
-%endif
+    %if ARCH_X86_64 == 0
+        %define r%1 e%1
+    %endif
 %endmacro

 DECLARE_REG_SIZE ax, al, ah
@@ -373,7 +393,7 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14

 %macro ASSERT 1
    %if (%1) == 0
-        %error assert failed
+        %error assertion ``%1'' failed
    %endif
 %endmacro

@@ -464,8 +484,10 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14
        %if %1 != 0 && required_stack_alignment > STACK_ALIGNMENT
            %if %1 > 0
                %assign regs_used (regs_used + 1)
-            %elif ARCH_X86_64 && regs_used == num_args && num_args <= 4 + UNIX64 * 2
-                %warning "Stack pointer will overwrite register argument"
+            %endif
+            %if ARCH_X86_64 && regs_used < 5 + UNIX64 * 3
+                ; Ensure that we don't clobber any registers containing arguments
+                %assign regs_used 5 + UNIX64 * 3
            %endif
        %endif
    %endif
@@ -579,9 +601,9 @@ DECLARE_REG 14, R15, 120
 %macro RET 0
    WIN64_RESTORE_XMM_INTERNAL rsp
    POP_IF_USED 14, 13, 12, 11, 10, 9, 8, 7
-%if mmsize == 32
-    vzeroupper
-%endif
+    %if mmsize == 32
+        vzeroupper
+    %endif
    AUTO_REP_RET
 %endmacro

@@ -618,17 +640,17 @@ DECLARE_REG 14, R15, 72
 %define has_epilogue regs_used > 9 || mmsize == 32 || stack_size > 0

 %macro RET 0
-%if stack_size_padded > 0
-%if required_stack_alignment > STACK_ALIGNMENT
-    mov rsp, rstkm
-%else
-    add rsp, stack_size_padded
-%endif
-%endif
+    %if stack_size_padded > 0
+        %if required_stack_alignment > STACK_ALIGNMENT
+            mov rsp, rstkm
+        %else
+            add rsp, stack_size_padded
+        %endif
+    %endif
    POP_IF_USED 14, 13, 12, 11, 10, 9
-%if mmsize == 32
-    vzeroupper
-%endif
+    %if mmsize == 32
+        vzeroupper
+    %endif
    AUTO_REP_RET
 %endmacro

@@ -674,29 +696,29 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
 %define has_epilogue regs_used > 3 || mmsize == 32 || stack_size > 0

 %macro RET 0
-%if stack_size_padded > 0
-%if required_stack_alignment > STACK_ALIGNMENT
-    mov rsp, rstkm
-%else
-    add rsp, stack_size_padded
-%endif
-%endif
+    %if stack_size_padded > 0
+        %if required_stack_alignment > STACK_ALIGNMENT
+            mov rsp, rstkm
+        %else
+            add rsp, stack_size_padded
+        %endif
+    %endif
    POP_IF_USED 6, 5, 4, 3
-%if mmsize == 32
-    vzeroupper
-%endif
+    %if mmsize == 32
+        vzeroupper
+    %endif
    AUTO_REP_RET
 %endmacro

 %endif ;======================================================================

 %if WIN64 == 0
-%macro WIN64_SPILL_XMM 1
-%endmacro
-%macro WIN64_RESTORE_XMM 1
-%endmacro
-%macro WIN64_PUSH_XMM 0
-%endmacro
+    %macro WIN64_SPILL_XMM 1
+    %endmacro
+    %macro WIN64_RESTORE_XMM 1
+    %endmacro
+    %macro WIN64_PUSH_XMM 0
+    %endmacro
 %endif

 ; On AMD cpus <=K10, an ordinary ret is slow if it immediately follows either
@@ -709,24 +731,26 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
    %else
        rep ret
    %endif
+    annotate_function_size
 %endmacro

 %define last_branch_adr $$
 %macro AUTO_REP_RET 0
-    %ifndef cpuflags
-        times ((last_branch_adr-$)>>31)+1 rep ; times 1 iff $ != last_branch_adr.
-    %elif notcpuflag(ssse3)
-        times ((last_branch_adr-$)>>31)+1 rep
+    %if notcpuflag(ssse3)
+        times ((last_branch_adr-$)>>31)+1 rep ; times 1 iff $ == last_branch_adr.
    %endif
    ret
+    annotate_function_size
 %endmacro

 %macro BRANCH_INSTR 0-*
    %rep %0
        %macro %1 1-2 %1
            %2 %1
-            %%branch_instr:
-            %xdefine last_branch_adr %%branch_instr
+            %if notcpuflag(ssse3)
+                %%branch_instr equ $
+                %xdefine last_branch_adr %%branch_instr
+            %endif
        %endmacro
        %rotate 1
    %endrep
@@ -741,6 +765,7 @@ BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg, jge, jng, jnge, ja, jae,
    %elif %2
        jmp %1
    %endif
+    annotate_function_size
 %endmacro

 ;=============================================================================
@@ -762,6 +787,7 @@ BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg, jge, jng, jnge, ja, jae,
    cglobal_internal 0, %1 %+ SUFFIX, %2
 %endmacro
 %macro cglobal_internal 2-3+
+    annotate_function_size
    %if %1
        %xdefine %%FUNCTION_PREFIX private_prefix
        ; libvpx explicitly sets visibility in shared object builds. Avoid
@@ -782,17 +808,10 @@ BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg, jge, jng, jnge, ja, jae,
        CAT_XDEFINE cglobaled_, %2, 1
    %endif
    %xdefine current_function %2
-    %ifidn __OUTPUT_FORMAT__,elf32
+    %xdefine current_function_section __SECT__
+    %if FORMAT_ELF
        global %2:function %%VISIBILITY
-    %elifidn __OUTPUT_FORMAT__,elf64
-        global %2:function %%VISIBILITY
-    %elifidn __OUTPUT_FORMAT__,macho32
-        %ifdef __NASM_VER__
-            global %2
-        %else
-            global %2:private_extern
-        %endif
-    %elifidn __OUTPUT_FORMAT__,macho64
+    %elif FORMAT_MACHO
        %ifdef __NASM_VER__
            global %2
        %else
@@ -822,16 +841,16 @@ BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg, jge, jng, jnge, ja, jae,

 ; like cextern, but without the prefix
 %macro cextern_naked 1
-    %xdefine %1 mangle(%1)
+    %ifdef PREFIX
+        %xdefine %1 mangle(%1)
+    %endif
    CAT_XDEFINE cglobaled_, %1, 1
    extern %1
 %endmacro

 %macro const 1-2+
    %xdefine %1 mangle(private_prefix %+ _ %+ %1)
-    %ifidn __OUTPUT_FORMAT__,elf32
-        global %1:data hidden
-    %elifidn __OUTPUT_FORMAT__,elf64
+    %if FORMAT_ELF
        global %1:data hidden
    %else
        global %1
@@ -839,14 +858,29 @@ BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg, jge, jng, jnge, ja, jae,
    %1: %2
 %endmacro

-; This is needed for ELF, otherwise the GNU linker assumes the stack is
-; executable by default.
-%ifidn __OUTPUT_FORMAT__,elf32
-SECTION .note.GNU-stack noalloc noexec nowrite progbits
-%elifidn __OUTPUT_FORMAT__,elf64
-SECTION .note.GNU-stack noalloc noexec nowrite progbits
+; This is needed for ELF, otherwise the GNU linker assumes the stack is executable by default.
+%if FORMAT_ELF
+    [SECTION .note.GNU-stack noalloc noexec nowrite progbits]
 %endif

+; Tell debuggers how large the function was.
+; This may be invoked multiple times per function; we rely on later instances overriding earlier ones.
+; This is invoked by RET and similar macros, and also cglobal does it for the previous function,
+; but if the last function in a source file doesn't use any of the standard macros for its epilogue,
+; then its size might be unspecified.
+%macro annotate_function_size 0
+    %ifdef __YASM_VER__
+        %ifdef current_function
+            %if FORMAT_ELF
+                current_function_section
+                %%ecf equ $
+                size current_function %%ecf - current_function
+                __SECT__
+            %endif
+        %endif
+    %endif
+%endmacro
+
 ; cpuflags

 %assign cpuflags_mmx      (1<<0)
@@ -875,12 +909,9 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
 %assign cpuflags_bmi1     (1<<22)|cpuflags_lzcnt
 %assign cpuflags_bmi2     (1<<23)|cpuflags_bmi1

-%define    cpuflag(x) ((cpuflags & (cpuflags_ %+ x)) == (cpuflags_ %+ x))
-%define notcpuflag(x) ((cpuflags & (cpuflags_ %+ x)) != (cpuflags_ %+ x))
-
-%ifdef __NASM_VER__
-    %use smartalign
-%endif
+; Returns a boolean value expressing whether or not the specified cpuflag is enabled.
+%define    cpuflag(x) (((((cpuflags & (cpuflags_ %+ x)) ^ (cpuflags_ %+ x)) - 1) >> 31) & 1)
+%define notcpuflag(x) (cpuflag(x) ^ 1)

 ; Takes an arbitrary number of cpuflags from the above list.
 ; All subsequent functions (up to the next INIT_CPUFLAGS) is built for the specified cpu.
@@ -917,12 +948,18 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
        %endif
    %endif

-    %ifdef __NASM_VER__
-        ALIGNMODE k7
-    %elif ARCH_X86_64 || cpuflag(sse2)
-        CPU amdnop
+    %if ARCH_X86_64 || cpuflag(sse2)
+        %ifdef __NASM_VER__
+            ALIGNMODE k8
+        %else
+            CPU amdnop
+        %endif
    %else
-        CPU basicnop
+        %ifdef __NASM_VER__
+            ALIGNMODE nop
+        %else
+            CPU basicnop
+        %endif
    %endif
 %endmacro

@@ -951,14 +988,14 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
    %define movnta movntq
    %assign %%i 0
    %rep 8
-    CAT_XDEFINE m, %%i, mm %+ %%i
-    CAT_XDEFINE nnmm, %%i, %%i
-    %assign %%i %%i+1
+        CAT_XDEFINE m, %%i, mm %+ %%i
+        CAT_XDEFINE nnmm, %%i, %%i
+        %assign %%i %%i+1
    %endrep
    %rep 8
-    CAT_UNDEF m, %%i
-    CAT_UNDEF nnmm, %%i
-    %assign %%i %%i+1
+        CAT_UNDEF m, %%i
+        CAT_UNDEF nnmm, %%i
+        %assign %%i %%i+1
    %endrep
    INIT_CPUFLAGS %1
 %endmacro
@@ -969,7 +1006,7 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
    %define mmsize 16
    %define num_mmregs 8
    %if ARCH_X86_64
-    %define num_mmregs 16
+        %define num_mmregs 16
    %endif
    %define mova movdqa
    %define movu movdqu
@@ -977,9 +1014,9 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
    %define movnta movntdq
    %assign %%i 0
    %rep num_mmregs
-    CAT_XDEFINE m, %%i, xmm %+ %%i
-    CAT_XDEFINE nnxmm, %%i, %%i
-    %assign %%i %%i+1
+        CAT_XDEFINE m, %%i, xmm %+ %%i
+        CAT_XDEFINE nnxmm, %%i, %%i
+        %assign %%i %%i+1
    %endrep
    INIT_CPUFLAGS %1
 %endmacro
@@ -990,7 +1027,7 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
    %define mmsize 32
    %define num_mmregs 8
    %if ARCH_X86_64
-    %define num_mmregs 16
+        %define num_mmregs 16
    %endif
    %define mova movdqa
    %define movu movdqu
@@ -998,9 +1035,9 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
    %define movnta movntdq
    %assign %%i 0
    %rep num_mmregs
-    CAT_XDEFINE m, %%i, ymm %+ %%i
-    CAT_XDEFINE nnymm, %%i, %%i
-    %assign %%i %%i+1
+        CAT_XDEFINE m, %%i, ymm %+ %%i
+        CAT_XDEFINE nnymm, %%i, %%i
+        %assign %%i %%i+1
    %endrep
    INIT_CPUFLAGS %1
 %endmacro
@@ -1024,7 +1061,7 @@ INIT_XMM
 %assign i 0
 %rep 16
    DECLARE_MMCAST i
-%assign i i+1
+    %assign i i+1
 %endrep

 ; I often want to use macros that permute their arguments. e.g. there's no
@@ -1042,23 +1079,23 @@ INIT_XMM
 ; doesn't cost any cycles.

 %macro PERMUTE 2-* ; takes a list of pairs to swap
-%rep %0/2
-    %xdefine %%tmp%2 m%2
-    %rotate 2
-%endrep
-%rep %0/2
-    %xdefine m%1 %%tmp%2
-    CAT_XDEFINE nn, m%1, %1
-    %rotate 2
-%endrep
+    %rep %0/2
+        %xdefine %%tmp%2 m%2
+        %rotate 2
+    %endrep
+    %rep %0/2
+        %xdefine m%1 %%tmp%2
+        CAT_XDEFINE nn, m%1, %1
+        %rotate 2
+    %endrep
 %endmacro

 %macro SWAP 2+ ; swaps a single chain (sometimes more concise than pairs)
-%ifnum %1 ; SWAP 0, 1, ...
-    SWAP_INTERNAL_NUM %1, %2
-%else ; SWAP m0, m1, ...
-    SWAP_INTERNAL_NAME %1, %2
-%endif
+    %ifnum %1 ; SWAP 0, 1, ...
+        SWAP_INTERNAL_NUM %1, %2
+    %else ; SWAP m0, m1, ...
+        SWAP_INTERNAL_NAME %1, %2
+    %endif
 %endmacro

 %macro SWAP_INTERNAL_NUM 2-*
@@ -1068,7 +1105,7 @@ INIT_XMM
        %xdefine m%2 %%tmp
        CAT_XDEFINE nn, m%1, %1
        CAT_XDEFINE nn, m%2, %2
-    %rotate 1
+        %rotate 1
    %endrep
 %endmacro

@@ -1076,7 +1113,7 @@ INIT_XMM
    %xdefine %%args nn %+ %1
    %rep %0-1
        %xdefine %%args %%args, nn %+ %2
-    %rotate 1
+        %rotate 1
    %endrep
    SWAP_INTERNAL_NUM %%args
 %endmacro
@@ -1093,7 +1130,7 @@ INIT_XMM
    %assign %%i 0
    %rep num_mmregs
        CAT_XDEFINE %%f, %%i, m %+ %%i
-    %assign %%i %%i+1
+        %assign %%i %%i+1
    %endrep
 %endmacro

@@ -1103,20 +1140,20 @@ INIT_XMM
        %rep num_mmregs
            CAT_XDEFINE m, %%i, %1_m %+ %%i
            CAT_XDEFINE nn, m %+ %%i, %%i
-        %assign %%i %%i+1
+            %assign %%i %%i+1
        %endrep
    %endif
 %endmacro

 ; Append cpuflags to the callee's name iff the appended name is known and the plain name isn't
 %macro call 1
-    call_internal %1, %1 %+ SUFFIX
+    call_internal %1 %+ SUFFIX, %1
 %endmacro
 %macro call_internal 2
-    %xdefine %%i %1
-    %ifndef cglobaled_%1
-        %ifdef cglobaled_%2
-            %xdefine %%i %2
+    %xdefine %%i %2
+    %ifndef cglobaled_%2
+        %ifdef cglobaled_%1
+            %xdefine %%i %1
        %endif
    %endif
    call %%i
@@ -1159,7 +1196,7 @@ INIT_XMM
    %endif
    CAT_XDEFINE sizeofxmm, i, 16
    CAT_XDEFINE sizeofymm, i, 32
-%assign i i+1
+    %assign i i+1
 %endrep
 %undef i

@@ -1536,7 +1573,7 @@ AVX_INSTR pfmul, 3dnow, 1, 0, 1
    %else
        CAT_XDEFINE q, j, i
    %endif
-%assign i i+1
+    %assign i i+1
 %endrep
 %undef i
 %undef j
@@ -1559,55 +1596,54 @@ FMA_INSTR  pmacsdd,  pmulld, paddd ; sse4 emulation
 FMA_INSTR pmacsdql,  pmuldq, paddq ; sse4 emulation
 FMA_INSTR pmadcswd, pmaddwd, paddd

-; convert FMA4 to FMA3 if possible
-%macro FMA4_INSTR 4
-    %macro %1 4-8 %1, %2, %3, %4
-        %if cpuflag(fma4)
-            v%5 %1, %2, %3, %4
-        %elifidn %1, %2
-            v%6 %1, %4, %3 ; %1 = %1 * %3 + %4
-        %elifidn %1, %3
-            v%7 %1, %2, %4 ; %1 = %2 * %1 + %4
-        %elifidn %1, %4
-            v%8 %1, %2, %3 ; %1 = %2 * %3 + %1
-        %else
-            %error fma3 emulation of ``%5 %1, %2, %3, %4'' is not supported
-        %endif
-    %endmacro
+; Macros for consolidating FMA3 and FMA4 using 4-operand (dst, src1, src2, src3) syntax.
+; FMA3 is only possible if dst is the same as one of the src registers.
+; Either src2 or src3 can be a memory operand.
+%macro FMA4_INSTR 2-*
+    %push fma4_instr
+    %xdefine %$prefix %1
+    %rep %0 - 1
+        %macro %$prefix%2 4-6 %$prefix, %2
+            %if notcpuflag(fma3) && notcpuflag(fma4)
+                %error use of ``%5%6'' fma instruction in cpuname function: current_function
+            %elif cpuflag(fma4)
+                v%5%6 %1, %2, %3, %4
+            %elifidn %1, %2
+                ; If %3 or %4 is a memory operand it needs to be encoded as the last operand.
+                %ifid %3
+                    v%{5}213%6 %2, %3, %4
+                %else
+                    v%{5}132%6 %2, %4, %3
+                %endif
+            %elifidn %1, %3
+                v%{5}213%6 %3, %2, %4
+            %elifidn %1, %4
+                v%{5}231%6 %4, %2, %3
+            %else
+                %error fma3 emulation of ``%5%6 %1, %2, %3, %4'' is not supported
+            %endif
+        %endmacro
+        %rotate 1
+    %endrep
+    %pop
 %endmacro

-FMA4_INSTR fmaddpd, fmadd132pd, fmadd213pd, fmadd231pd
-FMA4_INSTR fmaddps, fmadd132ps, fmadd213ps, fmadd231ps
-FMA4_INSTR fmaddsd, fmadd132sd, fmadd213sd, fmadd231sd
-FMA4_INSTR fmaddss, fmadd132ss, fmadd213ss, fmadd231ss
+FMA4_INSTR fmadd,    pd, ps, sd, ss
+FMA4_INSTR fmaddsub, pd, ps
+FMA4_INSTR fmsub,    pd, ps, sd, ss
+FMA4_INSTR fmsubadd, pd, ps
+FMA4_INSTR fnmadd,   pd, ps, sd, ss
+FMA4_INSTR fnmsub,   pd, ps, sd, ss

-FMA4_INSTR fmaddsubpd, fmaddsub132pd, fmaddsub213pd, fmaddsub231pd
-FMA4_INSTR fmaddsubps, fmaddsub132ps, fmaddsub213ps, fmaddsub231ps
-FMA4_INSTR fmsubaddpd, fmsubadd132pd, fmsubadd213pd, fmsubadd231pd
-FMA4_INSTR fmsubaddps, fmsubadd132ps, fmsubadd213ps, fmsubadd231ps
-
-FMA4_INSTR fmsubpd, fmsub132pd, fmsub213pd, fmsub231pd
-FMA4_INSTR fmsubps, fmsub132ps, fmsub213ps, fmsub231ps
-FMA4_INSTR fmsubsd, fmsub132sd, fmsub213sd, fmsub231sd
-FMA4_INSTR fmsubss, fmsub132ss, fmsub213ss, fmsub231ss
-
-FMA4_INSTR fnmaddpd, fnmadd132pd, fnmadd213pd, fnmadd231pd
-FMA4_INSTR fnmaddps, fnmadd132ps, fnmadd213ps, fnmadd231ps
-FMA4_INSTR fnmaddsd, fnmadd132sd, fnmadd213sd, fnmadd231sd
-FMA4_INSTR fnmaddss, fnmadd132ss, fnmadd213ss, fnmadd231ss
-
-FMA4_INSTR fnmsubpd, fnmsub132pd, fnmsub213pd, fnmsub231pd
-FMA4_INSTR fnmsubps, fnmsub132ps, fnmsub213ps, fnmsub231ps
-FMA4_INSTR fnmsubsd, fnmsub132sd, fnmsub213sd, fnmsub231sd
-FMA4_INSTR fnmsubss, fnmsub132ss, fnmsub213ss, fnmsub231ss
-
-; workaround: vpbroadcastq is broken in x86_32 due to a yasm bug
-%if ARCH_X86_64 == 0
-%macro vpbroadcastq 2
-%if sizeof%1 == 16
-    movddup %1, %2
-%else
-    vbroadcastsd %1, %2
-%endif
-%endmacro
+; workaround: vpbroadcastq is broken in x86_32 due to a yasm bug (fixed in 1.3.0)
+%ifdef __YASM_VER__
+    %if __YASM_VERSION_ID__ < 0x01030000 && ARCH_X86_64 == 0
+        %macro vpbroadcastq 2
+            %if sizeof%1 == 16
+                movddup %1, %2
+            %else
+                vbroadcastsd %1, %2
+            %endif
+        %endmacro
+    %endif
 %endif
--- a/tools_common.c
+++ b/tools_common.c
@@ -16,11 +16,11 @@

 #include "./tools_common.h"

-#if CONFIG_VP8_ENCODER || CONFIG_VP9_ENCODER || CONFIG_VP10_ENCODER
+#if CONFIG_VP8_ENCODER || CONFIG_VP9_ENCODER
 #include "vpx/vp8cx.h"
 #endif

-#if CONFIG_VP8_DECODER || CONFIG_VP9_DECODER || CONFIG_VP10_DECODER
+#if CONFIG_VP8_DECODER || CONFIG_VP9_DECODER
 #include "vpx/vp8dx.h"
 #endif

@@ -133,10 +133,6 @@ int read_yuv_frame(struct VpxInputContext *input_ctx, vpx_image_t *yuv_frame) {
 #if CONFIG_ENCODERS

 static const VpxInterface vpx_encoders[] = {
-#if CONFIG_VP10_ENCODER
-  {"vp10", VP10_FOURCC, &vpx_codec_vp10_cx},
-#endif
-
 #if CONFIG_VP8_ENCODER
  {"vp8", VP8_FOURCC, &vpx_codec_vp8_cx},
 #endif
@@ -178,10 +174,6 @@ static const VpxInterface vpx_decoders[] = {
 #if CONFIG_VP9_DECODER
  {"vp9", VP9_FOURCC, &vpx_codec_vp9_dx},
 #endif
-
-#if CONFIG_VP10_DECODER
-  {"vp10", VP10_FOURCC, &vpx_codec_vp10_dx},
-#endif
 };

 int get_vpx_decoder_count(void) {
--- a/tools_common.h
+++ b/tools_common.h
@@ -62,7 +62,6 @@

 #define VP8_FOURCC 0x30385056
 #define VP9_FOURCC 0x30395056
-#define VP10_FOURCC 0x303a5056

 enum VideoFileType {
  FILE_TYPE_RAW,
--- a/Show More
+++ b/Show More