@scrypted/libav 1.0.180 → 1.0.181

package/openh264/codec/common/loongarch/satd_sad_lasx.c

@@ -0,0 +1,568 @@
+/*!
+ * \copy
+ *     Copyright (c)  2009-2018, Cisco Systems
+ *     All rights reserved.
+ *
+ *     Redistribution and use in source and binary forms, with or without
+ *     modification, are permitted provided that the following conditions
+ *     are met:
+ *
+ *        * Redistributions of source code must retain the above copyright
+ *          notice, this list of conditions and the following disclaimer.
+ *
+ *        * Redistributions in binary form must reproduce the above copyright
+ *          notice, this list of conditions and the following disclaimer in
+ *          the documentation and/or other materials provided with the
+ *          distribution.
+ *
+ *     THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *     "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *     LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+ *     FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+ *     COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+ *     INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
+ *     BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ *     LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+ *     CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ *     LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
+ *     ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ *     POSSIBILITY OF SUCH DAMAGE.
+ *
+ *
+ * \file    satd_sad_lasx.c
+ *
+ * \brief   Loongson optimization
+ *
+ * \date    12/10/2021 Created
+ *
+ *************************************************************************************
+ */
+
+#include <stdint.h>
+#include "loongson_intrinsics.h"
+
+#define HORISUM(in0, in1, out0)            \
+  out0 = __lasx_xvabsd_bu(in0, in1);       \
+  out0 = __lasx_xvhaddw_hu_bu(out0, out0); \
+  out0 = __lasx_xvhaddw_wu_hu(out0, out0); \
+  out0 = __lasx_xvhaddw_du_wu(out0, out0); \
+
+int32_t WelsSampleSad4x4_lasx (uint8_t* pSample1, int32_t iStride1,
+                               uint8_t* pSample2, int32_t iStride2) {
+  uint8_t *pSrc1 = pSample1;
+  uint8_t *pSrc2 = pSample2;
+  int32_t iStride0 = 0;
+  int32_t iStride1_tmp = iStride1 << 1;
+  int32_t iStride2_tmp = iStride2 << 1;
+
+  __m256i src1_0, src1_1, src1_2, src1_3;
+  __m256i src2_0, src2_1, src2_2, src2_3;
+
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride0,
+            pSrc1, iStride1,
+            pSrc1, iStride1_tmp,
+            pSrc1, iStride1_tmp + iStride1,
+            src1_0, src1_1, src1_2, src1_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc2, iStride0,
+            pSrc2, iStride2,
+            pSrc2, iStride2_tmp,
+            pSrc2, iStride2_tmp + iStride2,
+            src2_0, src2_1, src2_2, src2_3);
+
+  DUP2_ARG2(__lasx_xvpackev_w,
+            src1_0, src1_1, src1_2, src1_3,
+            src1_0, src1_2);
+  DUP2_ARG2(__lasx_xvpackev_w,
+            src2_0, src2_1, src2_2, src2_3,
+            src2_0, src2_2);
+  DUP2_ARG2(__lasx_xvpackev_d,
+            src1_0, src1_2, src2_0, src2_2,
+            src1_0, src2_0);
+
+  HORISUM(src1_0, src2_0, src1_0);
+
+  src1_0 = __lasx_xvhaddw_qu_du(src1_0, src1_0);
+
+  return __lasx_xvpickve2gr_d(src1_0, 0);
+}
+
+static inline
+int32_t WelsSampleSad8x8x2_lasx (uint8_t* pSample1, int32_t iStride1,
+                                 uint8_t* pSample2, int32_t iStride2) {
+  uint8_t* pSrc1 = pSample1;
+  uint8_t* pSrc2 = pSample2;
+  int32_t iStride0 = 0;
+  int32_t iStride1_tmp2 = iStride1 << 1;
+  int32_t iStride1_tmp3 = iStride1_tmp2 + iStride1;
+  int32_t iStride1_tmp4 = iStride1 << 2;
+  int32_t iStride1_tmp5 = iStride1_tmp4 + iStride1;
+  int32_t iStride1_tmp6 = iStride1_tmp5 + iStride1;
+  int32_t iStride1_tmp7 = iStride1_tmp6 + iStride1;
+  int32_t iStride2_tmp2 = iStride2 << 1;
+  int32_t iStride2_tmp3 = iStride2_tmp2 + iStride2;
+  int32_t iStride2_tmp4 = iStride2 << 2;
+  int32_t iStride2_tmp5 = iStride2_tmp4 + iStride2;
+  int32_t iStride2_tmp6 = iStride2_tmp5 + iStride2;
+  int32_t iStride2_tmp7 = iStride2_tmp6 + iStride2;
+
+  __m256i src1_0, src1_1, src1_2, src1_3,
+          src1_4, src1_5, src1_6, src1_7;
+  __m256i src2_0, src2_1, src2_2, src2_3,
+          src2_4, src2_5, src2_6, src2_7;
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride0,
+            pSrc1, iStride1,
+            pSrc1, iStride1_tmp2,
+            pSrc1, iStride1_tmp3,
+            src1_0, src1_1, src1_2, src1_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride1_tmp4,
+            pSrc1, iStride1_tmp5,
+            pSrc1, iStride1_tmp6,
+            pSrc1, iStride1_tmp7,
+            src1_4, src1_5, src1_6, src1_7);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc2, iStride0,
+            pSrc2, iStride2,
+            pSrc2, iStride2_tmp2,
+            pSrc2, iStride2_tmp3,
+            src2_0, src2_1, src2_2, src2_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc2, iStride2_tmp4,
+            pSrc2, iStride2_tmp5,
+            pSrc2, iStride2_tmp6,
+            pSrc2, iStride2_tmp7,
+            src2_4, src2_5, src2_6, src2_7);
+  DUP4_ARG3(__lasx_xvpermi_q,
+            src1_0, src1_1, 0x20,
+            src1_2, src1_3, 0x20,
+            src1_4, src1_5, 0x20,
+            src1_6, src1_7, 0x20,
+            src1_0, src1_2, src1_4, src1_6);
+  DUP4_ARG3(__lasx_xvpermi_q,
+            src2_0, src2_1, 0x20,
+            src2_2, src2_3, 0x20,
+            src2_4, src2_5, 0x20,
+            src2_6, src2_7, 0x20,
+            src2_0, src2_2, src2_4, src2_6);
+  src1_0 = __lasx_xvabsd_bu(src1_0, src2_0);
+  src1_2 = __lasx_xvabsd_bu(src1_2, src2_2);
+  src1_4 = __lasx_xvabsd_bu(src1_4, src2_4);
+  src1_6 = __lasx_xvabsd_bu(src1_6, src2_6);
+  src1_0 = __lasx_xvhaddw_hu_bu(src1_0, src1_0);
+  src1_2 = __lasx_xvhaddw_hu_bu(src1_2, src1_2);
+  src1_4 = __lasx_xvhaddw_hu_bu(src1_4, src1_4);
+  src1_6 = __lasx_xvhaddw_hu_bu(src1_6, src1_6);
+  src1_0 = __lasx_xvadd_h(src1_0, src1_2);
+  src1_0 = __lasx_xvadd_h(src1_0, src1_4);
+  src1_0 = __lasx_xvadd_h(src1_0, src1_6);
+  src1_0 = __lasx_xvhaddw_wu_hu(src1_0, src1_0);
+  src1_0 = __lasx_xvhaddw_du_wu(src1_0, src1_0);
+  src1_0 = __lasx_xvhaddw_qu_du(src1_0, src1_0);
+  return (__lasx_xvpickve2gr_w(src1_0, 0) +
+          __lasx_xvpickve2gr_w(src1_0, 4));
+}
+
+int32_t WelsSampleSad8x8_lasx (uint8_t* pSample1, int32_t iStride1,
+                               uint8_t* pSample2, int32_t iStride2) {
+  uint8_t* pSrc1 = pSample1;
+  uint8_t* pSrc2 = pSample2;
+  int32_t iStride0 = 0;
+  int32_t iStride1_tmp2 = iStride1 << 1;
+  int32_t iStride1_tmp3 = iStride1_tmp2 + iStride1;
+  int32_t iStride1_tmp4 = iStride1 << 2;
+  int32_t iStride1_tmp5 = iStride1_tmp4 + iStride1;
+  int32_t iStride1_tmp6 = iStride1_tmp5 + iStride1;
+  int32_t iStride1_tmp7 = iStride1_tmp6 + iStride1;
+  int32_t iStride2_tmp2 = iStride2 << 1;
+  int32_t iStride2_tmp3 = iStride2_tmp2 + iStride2;
+  int32_t iStride2_tmp4 = iStride2 << 2;
+  int32_t iStride2_tmp5 = iStride2_tmp4 + iStride2;
+  int32_t iStride2_tmp6 = iStride2_tmp5 + iStride2;
+  int32_t iStride2_tmp7 = iStride2_tmp6 + iStride2;
+
+  __m256i src1_0, src1_1, src1_2, src1_3,
+          src1_4, src1_5, src1_6, src1_7;
+  __m256i src2_0, src2_1, src2_2, src2_3,
+          src2_4, src2_5, src2_6, src2_7;
+
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride0,
+            pSrc1, iStride1,
+            pSrc1, iStride1_tmp2,
+            pSrc1, iStride1_tmp3,
+            src1_0, src1_1, src1_2, src1_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride1_tmp4,
+            pSrc1, iStride1_tmp5,
+            pSrc1, iStride1_tmp6,
+            pSrc1, iStride1_tmp7,
+            src1_4, src1_5, src1_6, src1_7);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc2, iStride0,
+            pSrc2, iStride2,
+            pSrc2, iStride2_tmp2,
+            pSrc2, iStride2_tmp3,
+            src2_0, src2_1, src2_2, src2_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc2, iStride2_tmp4,
+            pSrc2, iStride2_tmp5,
+            pSrc2, iStride2_tmp6,
+            pSrc2, iStride2_tmp7,
+            src2_4, src2_5, src2_6, src2_7);
+
+  DUP4_ARG2(__lasx_xvpackev_d,
+            src1_0, src1_1, src1_2, src1_3,
+            src1_4, src1_5, src1_6, src1_7,
+            src1_0, src1_2, src1_4, src1_6);
+  DUP2_ARG3(__lasx_xvpermi_q,
+            src1_0, src1_2, 0x20,
+            src1_4, src1_6, 0x20,
+            src1_0, src1_4);
+  DUP4_ARG2(__lasx_xvpackev_d,
+            src2_0, src2_1, src2_2, src2_3,
+            src2_4, src2_5, src2_6, src2_7,
+            src2_0, src2_2, src2_4, src2_6);
+  DUP2_ARG3(__lasx_xvpermi_q,
+            src2_0, src2_2, 0x20,
+            src2_4, src2_6, 0x20,
+            src2_0, src2_4);
+
+  HORISUM(src1_0, src2_0, src1_0);
+  HORISUM(src1_4, src2_4, src1_4);
+
+  src1_0 = __lasx_xvadd_d(src1_0, src1_4);
+  src1_0 = __lasx_xvhaddw_qu_du(src1_0, src1_0);
+
+  return (__lasx_xvpickve2gr_d(src1_0, 0) +
+          __lasx_xvpickve2gr_d(src1_0, 2));
+}
+
+int32_t WelsSampleSatd4x4_lasx (uint8_t* pSample1, int32_t iStride1,
+                                uint8_t* pSample2, int32_t iStride2) {
+  int32_t iSatdSum;
+  uint8_t* pSrc1 = pSample1;
+  uint8_t* pSrc2 = pSample2;
+  int32_t iStride0 = 0;
+  int32_t iStride1_tmp = iStride1 << 1;
+  int32_t iStride2_tmp = iStride2 << 1;
+
+  __m256i src1_0, src1_1, src1_2, src1_3;
+  __m256i src2_0, src2_1, src2_2, src2_3;
+  __m256i iSample01, iSample23;
+  __m256i tmp0, tmp1, tmp2, tmp3;
+  __m256i zero = __lasx_xvldi(0);
+  v16i16 mask= {1, 0, 3, 2, 5, 4, 7, 6, 1, 0, 3, 2, 5, 4, 7, 6};
+
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride0,
+            pSrc1, iStride1,
+            pSrc1, iStride1_tmp,
+            pSrc1, iStride1_tmp + iStride1,
+            src1_0, src1_1, src1_2, src1_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc2, iStride0,
+            pSrc2, iStride2,
+            pSrc2, iStride2_tmp,
+            pSrc2, iStride2_tmp + iStride2,
+            src2_0, src2_1, src2_2, src2_3);
+  DUP4_ARG2(__lasx_xvpackev_w,
+            src1_0, src1_1,
+            src1_2, src1_3,
+            src2_0, src2_1,
+            src2_2, src2_3,
+            src1_0, src1_2, src2_0, src2_2);
+  DUP2_ARG2(__lasx_xvpackev_d,
+            src1_0, src1_2,
+            src2_0, src2_2,
+            src1_0, src2_0);
+
+  tmp0 = __lasx_xvsubwev_h_bu(src1_0, src2_0);
+  tmp1 = __lasx_xvsubwod_h_bu(src1_0, src2_0);
+  tmp2 = __lasx_xvilvl_w(tmp0, tmp1);
+  tmp3 = __lasx_xvilvh_w(tmp0, tmp1);
+  tmp0 = __lasx_xvpermi_q(tmp3, tmp2, 0x20);
+  tmp0 = __lasx_xvshuf_h((__m256i)mask, tmp0, tmp0);
+
+  iSample01 = __lasx_xvhaddw_w_h(tmp0, tmp0);
+  iSample23 = __lasx_xvhsubw_w_h(tmp0, tmp0);
+  tmp0 = __lasx_xvhaddw_d_w(iSample01, iSample01);
+  tmp1 = __lasx_xvhaddw_d_w(iSample23, iSample23);
+  tmp2 = __lasx_xvhsubw_d_w(iSample23, iSample23);
+  tmp3 = __lasx_xvhsubw_d_w(iSample01, iSample01);
+
+  tmp1 = __lasx_xvpackev_w(tmp1, tmp0);
+  tmp3 = __lasx_xvpackev_w(tmp3, tmp2);
+  tmp0 = __lasx_xvpermi_q(tmp3, tmp1, 0x20);
+  tmp2 = __lasx_xvpermi_q(tmp3, tmp1, 0x31);
+  tmp0 = __lasx_xvpermi_w(tmp0, tmp0, 0x72);
+  tmp2 = __lasx_xvpermi_w(tmp2, tmp2, 0x72);
+
+  iSample01 = __lasx_xvadd_w(tmp0, tmp2);
+  iSample23 = __lasx_xvsub_w(tmp0, tmp2);
+
+  tmp0 = __lasx_xvhaddw_d_w(iSample01, iSample01);
+  tmp1 = __lasx_xvhaddw_d_w(iSample23, iSample23);
+  tmp2 = __lasx_xvhsubw_d_w(iSample23, iSample23);
+  tmp3 = __lasx_xvhsubw_d_w(iSample01, iSample01);
+
+  tmp0 = __lasx_xvpackev_w(tmp0, tmp1);
+  tmp2 = __lasx_xvpackev_w(tmp2, tmp3);
+
+  tmp0 = __lasx_xvabsd_w(tmp0, zero);
+  tmp2 = __lasx_xvabsd_w(tmp2, zero);
+  tmp0 = __lasx_xvadd_w(tmp0, tmp2);
+  tmp0 = __lasx_xvhaddw_d_w(tmp0, tmp0);
+  tmp0 = __lasx_xvhaddw_q_d(tmp0, tmp0);
+
+  iSatdSum = __lasx_xvpickve2gr_d(tmp0, 0) +
+             __lasx_xvpickve2gr_d(tmp0, 2);
+
+  return ((iSatdSum + 1) >> 1);
+}
+
+int32_t WelsSampleSad16x8_lasx (uint8_t* pSample1, int32_t iStride1,
+                                uint8_t* pSample2, int32_t iStride2) {
+
+  return WelsSampleSad8x8x2_lasx (pSample1, iStride1,
+                                  pSample2, iStride2);
+}
+
+int32_t WelsSampleSad8x16_lasx (uint8_t* pSample1, int32_t iStride1,
+                                uint8_t* pSample2, int32_t iStride2) {
+  int32_t iSadSum = 0;
+
+  iSadSum += WelsSampleSad8x8_lasx (pSample1, iStride1,
+                                    pSample2, iStride2);
+  iSadSum += WelsSampleSad8x8_lasx (pSample1 + (iStride1 << 3), iStride1,
+                                    pSample2 + (iStride2 << 3), iStride2);
+  return iSadSum;
+}
+
+int32_t WelsSampleSad16x16_lasx (uint8_t* pSample1, int32_t iStride1,
+                                 uint8_t* pSample2, int32_t iStride2) {
+  int32_t iSadSum = 0;
+
+  iSadSum += WelsSampleSad8x8x2_lasx (pSample1, iStride1,
+                                      pSample2, iStride2);
+  iSadSum += WelsSampleSad8x8x2_lasx (pSample1 + (iStride1 << 3), iStride1,
+                                      pSample2 + (iStride2 << 3), iStride2);
+  return iSadSum;
+}
+
+void WelsSampleSadFour4x4_lasx (uint8_t* iSample1, int32_t iStride1,
+                                uint8_t* iSample2, int32_t iStride2,
+                                int32_t* pSad) {
+  uint8_t *pSrc1 = iSample1;
+  uint8_t *pSrc2 = iSample2 - iStride2;
+  uint8_t *pSrc3 = iSample2 + iStride2;
+  uint8_t *pSrc4 = iSample2 - 1;
+  uint8_t *pSrc5 = iSample2 + 1;
+  int32_t iStride0 = 0;
+  int32_t iStride1_tmp = iStride1 << 1;
+  int32_t iStride2_tmp = iStride2 << 1;
+
+  __m256i src1_0, src1_1, src1_2, src1_3;
+  __m256i src2_0, src2_1, src2_2, src2_3;
+  __m256i cb0, cb1, cb2, cb3, cb4, cb5, cb6, cb7;
+
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride0,
+            pSrc1, iStride1,
+            pSrc1, iStride1_tmp,
+            pSrc1, iStride1_tmp + iStride1,
+            src1_0, src1_1, src1_2, src1_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc2, iStride0,
+            pSrc2, iStride2,
+            pSrc2, iStride2_tmp,
+            pSrc2, iStride2_tmp + iStride2,
+            src2_0, src2_1, src2_2, src2_3);
+  DUP4_ARG2(__lasx_xvpackev_w,
+            src1_0, src1_1, src1_2, src1_3,
+            src2_0, src2_1, src2_2, src2_3,
+            src1_0, src1_2, src2_0, src2_2);
+  DUP2_ARG2(__lasx_xvpackev_d,
+            src1_0, src1_2, src2_0, src2_2,
+            cb0, cb1); //16 16
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride0,
+            pSrc1, iStride1,
+            pSrc1, iStride1_tmp,
+            pSrc1, iStride1_tmp + iStride1,
+            src1_0, src1_1, src1_2, src1_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc3, iStride0,
+            pSrc3, iStride2,
+            pSrc3, iStride2_tmp,
+            pSrc3, iStride2_tmp + iStride2,
+            src2_0, src2_1, src2_2, src2_3);
+  DUP4_ARG2(__lasx_xvpackev_w,
+            src1_0, src1_1, src1_2, src1_3,
+            src2_0, src2_1, src2_2, src2_3,
+            src1_0, src1_2, src2_0, src2_2);
+  DUP2_ARG2(__lasx_xvpackev_d,
+            src1_0, src1_2, src2_0, src2_2,
+            cb2, cb3); //16 16
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride0,
+            pSrc1, iStride1,
+            pSrc1, iStride1_tmp,
+            pSrc1, iStride1_tmp + iStride1,
+            src1_0, src1_1, src1_2, src1_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc4, iStride0,
+            pSrc4, iStride2,
+            pSrc4, iStride2_tmp,
+            pSrc4, iStride2_tmp + iStride2,
+            src2_0, src2_1, src2_2, src2_3);
+  DUP4_ARG2(__lasx_xvpackev_w,
+            src1_0, src1_1, src1_2, src1_3,
+            src2_0, src2_1, src2_2, src2_3,
+            src1_0, src1_2, src2_0, src2_2);
+  DUP2_ARG2(__lasx_xvpackev_d,
+            src1_0, src1_2, src2_0, src2_2,
+            cb4, cb5); //16 16
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc1, iStride0,
+            pSrc1, iStride1,
+            pSrc1, iStride1_tmp,
+            pSrc1, iStride1_tmp + iStride1,
+            src1_0, src1_1, src1_2, src1_3);
+  DUP4_ARG2(__lasx_xvldx,
+            pSrc5, iStride0,
+            pSrc5, iStride2,
+            pSrc5, iStride2_tmp,
+            pSrc5, iStride2_tmp + iStride2,
+            src2_0, src2_1, src2_2, src2_3);
+  DUP4_ARG2(__lasx_xvpackev_w,
+            src1_0, src1_1, src1_2, src1_3,
+            src2_0, src2_1, src2_2, src2_3,
+            src1_0, src1_2, src2_0, src2_2);
+  DUP2_ARG2(__lasx_xvpackev_d,
+            src1_0, src1_2, src2_0, src2_2,
+            cb6, cb7); //16 16
+
+  cb0 = __lasx_xvpermi_q(cb2, cb0, 0x20);
+  cb1 = __lasx_xvpermi_q(cb3, cb1, 0x20);
+  cb4 = __lasx_xvpermi_q(cb6, cb4, 0x20);
+  cb5 = __lasx_xvpermi_q(cb7, cb5, 0x20);
+
+  HORISUM(cb0, cb1, cb0);
+  HORISUM(cb4, cb5, cb4);
+
+  DUP2_ARG2(__lasx_xvhaddw_qu_du,
+           cb0, cb0, cb4, cb4,
+           cb0, cb4);
+
+  * (pSad) = __lasx_xvpickve2gr_d(cb0, 0);
+  * (pSad + 1) = __lasx_xvpickve2gr_d(cb0, 2);
+  * (pSad + 2) = __lasx_xvpickve2gr_d(cb4, 0);
+  * (pSad + 3) = __lasx_xvpickve2gr_d(cb4, 2);
+}
+
+void WelsSampleSadFour8x8_lasx (uint8_t* iSample1, int32_t iStride1,
+                                uint8_t* iSample2, int32_t iStride2,
+                                int32_t* pSad) {
+  * (pSad)     = WelsSampleSad8x8_lasx (iSample1, iStride1,
+                                       (iSample2 - iStride2), iStride2);
+  * (pSad + 1) = WelsSampleSad8x8_lasx (iSample1, iStride1,
+                                       (iSample2 + iStride2), iStride2);
+  * (pSad + 2) = WelsSampleSad8x8_lasx (iSample1, iStride1,
+                                       (iSample2 - 1), iStride2);
+  * (pSad + 3) = WelsSampleSad8x8_lasx (iSample1, iStride1,
+                                       (iSample2 + 1), iStride2);
+}
+
+void WelsSampleSadFour8x16_lasx (uint8_t* iSample1, int32_t iStride1,
+                                 uint8_t* iSample2, int32_t iStride2,
+                                 int32_t* pSad) {
+  * (pSad)     = WelsSampleSad8x16_lasx (iSample1, iStride1,
+                                        (iSample2 - iStride2), iStride2);
+  * (pSad + 1) = WelsSampleSad8x16_lasx (iSample1, iStride1,
+                                        (iSample2 + iStride2), iStride2);
+  * (pSad + 2) = WelsSampleSad8x16_lasx (iSample1, iStride1,
+                                        (iSample2 - 1), iStride2);
+  * (pSad + 3) = WelsSampleSad8x16_lasx (iSample1, iStride1,
+                                        (iSample2 + 1), iStride2);
+}
+
+void WelsSampleSadFour16x8_lasx (uint8_t* iSample1, int32_t iStride1,
+                                 uint8_t* iSample2, int32_t iStride2,
+                                 int32_t* pSad) {
+  * (pSad)     = WelsSampleSad16x8_lasx (iSample1, iStride1,
+                                        (iSample2 - iStride2), iStride2);
+  * (pSad + 1) = WelsSampleSad16x8_lasx (iSample1, iStride1,
+                                        (iSample2 + iStride2), iStride2);
+  * (pSad + 2) = WelsSampleSad16x8_lasx (iSample1, iStride1,
+                                        (iSample2 - 1), iStride2);
+  * (pSad + 3) = WelsSampleSad16x8_lasx (iSample1, iStride1,
+                                        (iSample2 + 1), iStride2);
+}
+
+void WelsSampleSadFour16x16_lasx (uint8_t* iSample1, int32_t iStride1,
+                                  uint8_t* iSample2, int32_t iStride2,
+                                  int32_t* pSad) {
+  * (pSad)     = WelsSampleSad16x16_lasx (iSample1, iStride1,
+                                         (iSample2 - iStride2), iStride2);
+  * (pSad + 1) = WelsSampleSad16x16_lasx (iSample1, iStride1,
+                                         (iSample2 + iStride2), iStride2);
+  * (pSad + 2) = WelsSampleSad16x16_lasx (iSample1, iStride1,
+                                         (iSample2 - 1), iStride2);
+  * (pSad + 3) = WelsSampleSad16x16_lasx (iSample1, iStride1,
+                                         (iSample2 + 1), iStride2);
+}
+
+int32_t WelsSampleSatd8x8_lasx (uint8_t* pSample1, int32_t iStride1,
+                                uint8_t* pSample2, int32_t iStride2) {
+  int32_t iSatdSum = 0;
+
+  iSatdSum += WelsSampleSatd4x4_lasx (pSample1, iStride1,
+                                      pSample2, iStride2);
+  iSatdSum += WelsSampleSatd4x4_lasx (pSample1 + 4, iStride1,
+                                      pSample2 + 4, iStride2);
+  iSatdSum += WelsSampleSatd4x4_lasx (pSample1 + (iStride1 << 2), iStride1,
+                                      pSample2 + (iStride2 << 2),   iStride2);
+  iSatdSum += WelsSampleSatd4x4_lasx (pSample1 + (iStride1 << 2) + 4, iStride1,
+                                      pSample2 + (iStride2 << 2) + 4, iStride2);
+  return iSatdSum;
+}
+
+int32_t WelsSampleSatd16x8_lasx (uint8_t* pSample1, int32_t iStride1,
+                                 uint8_t* pSample2, int32_t iStride2) {
+  int32_t iSatdSum = 0;
+
+  iSatdSum += WelsSampleSatd8x8_lasx (pSample1, iStride1,
+                                      pSample2, iStride2);
+  iSatdSum += WelsSampleSatd8x8_lasx (pSample1 + 8, iStride1,
+                                      pSample2 + 8, iStride2);
+  return iSatdSum;
+}
+
+int32_t WelsSampleSatd8x16_lasx (uint8_t* pSample1, int32_t iStride1,
+                                 uint8_t* pSample2, int32_t iStride2) {
+  int32_t iSatdSum = 0;
+
+  iSatdSum += WelsSampleSatd8x8_lasx (pSample1, iStride1,
+                                      pSample2, iStride2);
+  iSatdSum += WelsSampleSatd8x8_lasx (pSample1 + (iStride1 << 3), iStride1,
+                                      pSample2 + (iStride2 << 3), iStride2);
+  return iSatdSum;
+}
+
+int32_t WelsSampleSatd16x16_lasx (uint8_t* pSample1, int32_t iStride1,
+                                  uint8_t* pSample2, int32_t iStride2) {
+  int32_t iSatdSum = 0;
+
+  iSatdSum += WelsSampleSatd8x8_lasx (pSample1, iStride1,
+                                      pSample2, iStride2);
+  iSatdSum += WelsSampleSatd8x8_lasx (pSample1 + 8, iStride1,
+                                      pSample2 + 8, iStride2);
+  iSatdSum += WelsSampleSatd8x8_lasx (pSample1 + (iStride1 << 3), iStride1,
+                                      pSample2 + (iStride2 << 3), iStride2);
+  iSatdSum += WelsSampleSatd8x8_lasx (pSample1 + (iStride1 << 3) + 8, iStride1,
+                                      pSample2 + (iStride2 << 3) + 8, iStride2);
+  return iSatdSum;
+}

package/openh264/codec/common/meson.build

@@ -0,0 +1,84 @@
+cpp_sources = [
+  'src/common_tables.cpp',
+  'src/copy_mb.cpp',
+  'src/cpu.cpp',
+  'src/crt_util_safe_x.cpp',
+  'src/deblocking_common.cpp',
+  'src/expand_pic.cpp',
+  'src/intra_pred_common.cpp',
+  'src/mc.cpp',
+  'src/memory_align.cpp',
+  'src/sad_common.cpp',
+  'src/utils.cpp',
+  'src/welsCodecTrace.cpp',
+  'src/WelsTaskThread.cpp',
+  'src/WelsThread.cpp',
+  'src/WelsThreadLib.cpp',
+  'src/WelsThreadPool.cpp',
+]
+
+objs_asm = []
+if cpu_family in ['x86', 'x86_64']
+  asm_sources = [
+    'x86/cpuid.asm',
+    'x86/dct.asm',
+    'x86/deblock.asm',
+    'x86/expand_picture.asm',
+    'x86/intra_pred_com.asm',
+    'x86/mb_copy.asm',
+    'x86/mc_chroma.asm',
+    'x86/mc_luma.asm',
+    'x86/satd_sad.asm',
+    'x86/vaa.asm',
+  ]
+  objs_asm += asm_gen.process(asm_sources)
+elif cpu_family == 'arm'
+  asm_sources = [
+    'arm/copy_mb_neon.S',
+    'arm/deblocking_neon.S',
+    'arm/expand_picture_neon.S',
+    'arm/intra_pred_common_neon.S',
+    'arm/mc_neon.S',
+  ]
+  if use_asm_gen
+    objs_asm = asm_gen.process(asm_sources)
+  else
+    cpp_sources += asm_sources
+  endif
+elif cpu_family == 'aarch64'
+  asm_sources = [
+    'arm64/copy_mb_aarch64_neon.S',
+    'arm64/deblocking_aarch64_neon.S',
+    'arm64/expand_picture_aarch64_neon.S',
+    'arm64/intra_pred_common_aarch64_neon.S',
+    'arm64/mc_aarch64_neon.S',
+  ]
+  if use_asm_gen
+    objs_asm = asm_gen.process(asm_sources)
+  else
+    cpp_sources += asm_sources
+  endif
+elif cpu_family in ['loongarch32', 'loongarch64']
+  asm_sources = [
+    'loongarch/copy_mb_lsx.c',
+    'loongarch/deblock_lsx.c',
+    'loongarch/intra_pred_com_lsx.c',
+    'loongarch/intra_pred_com_lasx.c',
+    'loongarch/mc_chroma_lsx.c',
+    'loongarch/mc_horver_lsx.c',
+    'loongarch/satd_sad_lasx.c',
+  ]
+  cpp_sources += asm_sources
+elif cpu_family == 'riscv64'
+  asm_sources = []
+elif cpu_family == 'ppc'
+  asm_sources = []
+elif cpu_family == 'ppc64'
+  asm_sources = []
+else
+  error('Unsupported cpu_family @0@'.format(cpu_family))
+endif
+
+libcommon = static_library('common', cpp_sources, objs_asm,
+  include_directories: [inc, casm_inc],
+  dependencies: deps)