avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for qpel functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for qpel functions in new file qpeldsp_msa.c
Adds new generic macros (needed for this patch) in libavutil/mips/generic_macros_msa.h

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/mips/Makefile

@@ -25,6 +25,7 @@ OBJS-$(CONFIG_H264QPEL)                   += mips/h264qpel_init_mips.o
 OBJS-$(CONFIG_H264CHROMA)                 += mips/h264chroma_init_mips.o
 OBJS-$(CONFIG_H264PRED)                   += mips/h264pred_init_mips.o
 OBJS-$(CONFIG_H263DSP)                    += mips/h263dsp_init_mips.o
+OBJS-$(CONFIG_QPELDSP)                    += mips/qpeldsp_init_mips.o
 MSA-OBJS-$(CONFIG_HEVC_DECODER)           += mips/hevcdsp_msa.o            \
                                              mips/hevc_mc_uni_msa.o        \
                                              mips/hevc_mc_uniw_msa.o       \
@@ -39,5 +40,6 @@ MSA-OBJS-$(CONFIG_H264QPEL)               += mips/h264qpel_msa.o
 MSA-OBJS-$(CONFIG_H264CHROMA)             += mips/h264chroma_msa.o
 MSA-OBJS-$(CONFIG_H264PRED)               += mips/h264pred_msa.o
 MSA-OBJS-$(CONFIG_H263DSP)                += mips/h263dsp_msa.o
+MSA-OBJS-$(CONFIG_QPELDSP)                += mips/qpeldsp_msa.o
 LOONGSON3-OBJS-$(CONFIG_H264DSP)          += mips/h264dsp_mmi.o
 LOONGSON3-OBJS-$(CONFIG_H264CHROMA)       += mips/h264chroma_mmi.o

libavcodec/mips/qpeldsp_init_mips.c

@@ -0,0 +1,167 @@
+/*
+ * Copyright (c) 2015 Parag Salasakar (Parag.Salasakar@imgtec.com)
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "qpeldsp_mips.h"
+
+#if HAVE_MSA
+static av_cold void qpeldsp_init_msa(QpelDSPContext *c)
+{
+    c->put_qpel_pixels_tab[0][0] = ff_copy_16x16_msa;
+    c->put_qpel_pixels_tab[0][1] = ff_horiz_mc_qpel_aver_src0_16width_msa;
+    c->put_qpel_pixels_tab[0][2] = ff_horiz_mc_qpel_16width_msa;
+    c->put_qpel_pixels_tab[0][3] = ff_horiz_mc_qpel_aver_src1_16width_msa;
+    c->put_qpel_pixels_tab[0][4] = ff_vert_mc_qpel_aver_src0_16x16_msa;
+    c->put_qpel_pixels_tab[0][5] = ff_hv_mc_qpel_aver_hv_src00_16x16_msa;
+    c->put_qpel_pixels_tab[0][6] = ff_hv_mc_qpel_aver_v_src0_16x16_msa;
+    c->put_qpel_pixels_tab[0][7] = ff_hv_mc_qpel_aver_hv_src10_16x16_msa;
+    c->put_qpel_pixels_tab[0][8] = ff_vert_mc_qpel_16x16_msa;
+    c->put_qpel_pixels_tab[0][9] = ff_hv_mc_qpel_aver_h_src0_16x16_msa;
+    c->put_qpel_pixels_tab[0][10] = ff_hv_mc_qpel_16x16_msa;
+    c->put_qpel_pixels_tab[0][11] = ff_hv_mc_qpel_aver_h_src1_16x16_msa;
+    c->put_qpel_pixels_tab[0][12] = ff_vert_mc_qpel_aver_src1_16x16_msa;
+    c->put_qpel_pixels_tab[0][13] = ff_hv_mc_qpel_aver_hv_src01_16x16_msa;
+    c->put_qpel_pixels_tab[0][14] = ff_hv_mc_qpel_aver_v_src1_16x16_msa;
+    c->put_qpel_pixels_tab[0][15] = ff_hv_mc_qpel_aver_hv_src11_16x16_msa;
+
+    c->put_qpel_pixels_tab[1][0] = ff_copy_8x8_msa;
+    c->put_qpel_pixels_tab[1][1] = ff_horiz_mc_qpel_aver_src0_8width_msa;
+    c->put_qpel_pixels_tab[1][2] = ff_horiz_mc_qpel_8width_msa;
+    c->put_qpel_pixels_tab[1][3] = ff_horiz_mc_qpel_aver_src1_8width_msa;
+    c->put_qpel_pixels_tab[1][4] = ff_vert_mc_qpel_aver_src0_8x8_msa;
+    c->put_qpel_pixels_tab[1][5] = ff_hv_mc_qpel_aver_hv_src00_8x8_msa;
+    c->put_qpel_pixels_tab[1][6] = ff_hv_mc_qpel_aver_v_src0_8x8_msa;
+    c->put_qpel_pixels_tab[1][7] = ff_hv_mc_qpel_aver_hv_src10_8x8_msa;
+    c->put_qpel_pixels_tab[1][8] = ff_vert_mc_qpel_8x8_msa;
+    c->put_qpel_pixels_tab[1][9] = ff_hv_mc_qpel_aver_h_src0_8x8_msa;
+    c->put_qpel_pixels_tab[1][10] = ff_hv_mc_qpel_8x8_msa;
+    c->put_qpel_pixels_tab[1][11] = ff_hv_mc_qpel_aver_h_src1_8x8_msa;
+    c->put_qpel_pixels_tab[1][12] = ff_vert_mc_qpel_aver_src1_8x8_msa;
+    c->put_qpel_pixels_tab[1][13] = ff_hv_mc_qpel_aver_hv_src01_8x8_msa;
+    c->put_qpel_pixels_tab[1][14] = ff_hv_mc_qpel_aver_v_src1_8x8_msa;
+    c->put_qpel_pixels_tab[1][15] = ff_hv_mc_qpel_aver_hv_src11_8x8_msa;
+
+    c->put_no_rnd_qpel_pixels_tab[0][0] = ff_copy_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][1] =
+        ff_horiz_mc_qpel_no_rnd_aver_src0_16width_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][2] = ff_horiz_mc_qpel_no_rnd_16width_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][3] =
+        ff_horiz_mc_qpel_no_rnd_aver_src1_16width_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][4] =
+        ff_vert_mc_qpel_no_rnd_aver_src0_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][5] =
+        ff_hv_mc_qpel_no_rnd_aver_hv_src00_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][6] =
+        ff_hv_mc_qpel_no_rnd_aver_v_src0_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][7] =
+        ff_hv_mc_qpel_no_rnd_aver_hv_src10_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][8] = ff_vert_mc_qpel_no_rnd_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][9] =
+        ff_hv_mc_qpel_no_rnd_aver_h_src0_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][10] = ff_hv_mc_qpel_no_rnd_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][11] =
+        ff_hv_mc_qpel_no_rnd_aver_h_src1_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][12] =
+        ff_vert_mc_qpel_no_rnd_aver_src1_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][13] =
+        ff_hv_mc_qpel_no_rnd_aver_hv_src01_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][14] =
+        ff_hv_mc_qpel_no_rnd_aver_v_src1_16x16_msa;
+    c->put_no_rnd_qpel_pixels_tab[0][15] =
+        ff_hv_mc_qpel_no_rnd_aver_hv_src11_16x16_msa;
+
+    c->put_no_rnd_qpel_pixels_tab[1][0] = ff_copy_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][1] =
+        ff_horiz_mc_qpel_no_rnd_aver_src0_8width_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][2] = ff_horiz_mc_qpel_no_rnd_8width_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][3] =
+        ff_horiz_mc_qpel_no_rnd_aver_src1_8width_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][4] =
+        ff_vert_mc_qpel_no_rnd_aver_src0_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][5] =
+        ff_hv_mc_qpel_no_rnd_aver_hv_src00_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][6] =
+        ff_hv_mc_qpel_no_rnd_aver_v_src0_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][7] =
+        ff_hv_mc_qpel_no_rnd_aver_hv_src10_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][8] = ff_vert_mc_qpel_no_rnd_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][9] =
+        ff_hv_mc_qpel_no_rnd_aver_h_src0_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][10] = ff_hv_mc_qpel_no_rnd_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][11] =
+        ff_hv_mc_qpel_no_rnd_aver_h_src1_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][12] =
+        ff_vert_mc_qpel_no_rnd_aver_src1_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][13] =
+        ff_hv_mc_qpel_no_rnd_aver_hv_src01_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][14] =
+        ff_hv_mc_qpel_no_rnd_aver_v_src1_8x8_msa;
+    c->put_no_rnd_qpel_pixels_tab[1][15] =
+        ff_hv_mc_qpel_no_rnd_aver_hv_src11_8x8_msa;
+
+    c->avg_qpel_pixels_tab[0][0] = ff_avg_width16_msa;
+    c->avg_qpel_pixels_tab[0][1] =
+        ff_horiz_mc_qpel_avg_dst_aver_src0_16width_msa;
+    c->avg_qpel_pixels_tab[0][2] = ff_horiz_mc_qpel_avg_dst_16width_msa;
+    c->avg_qpel_pixels_tab[0][3] =
+        ff_horiz_mc_qpel_avg_dst_aver_src1_16width_msa;
+    c->avg_qpel_pixels_tab[0][4] = ff_vert_mc_qpel_avg_dst_aver_src0_16x16_msa;
+    c->avg_qpel_pixels_tab[0][5] =
+        ff_hv_mc_qpel_avg_dst_aver_hv_src00_16x16_msa;
+    c->avg_qpel_pixels_tab[0][6] = ff_hv_mc_qpel_avg_dst_aver_v_src0_16x16_msa;
+    c->avg_qpel_pixels_tab[0][7] =
+        ff_hv_mc_qpel_avg_dst_aver_hv_src10_16x16_msa;
+    c->avg_qpel_pixels_tab[0][8] = ff_vert_mc_qpel_avg_dst_16x16_msa;
+    c->avg_qpel_pixels_tab[0][9] = ff_hv_mc_qpel_avg_dst_aver_h_src0_16x16_msa;
+    c->avg_qpel_pixels_tab[0][10] = ff_hv_mc_qpel_avg_dst_16x16_msa;
+    c->avg_qpel_pixels_tab[0][11] = ff_hv_mc_qpel_avg_dst_aver_h_src1_16x16_msa;
+    c->avg_qpel_pixels_tab[0][12] = ff_vert_mc_qpel_avg_dst_aver_src1_16x16_msa;
+    c->avg_qpel_pixels_tab[0][13] =
+        ff_hv_mc_qpel_avg_dst_aver_hv_src01_16x16_msa;
+    c->avg_qpel_pixels_tab[0][14] = ff_hv_mc_qpel_avg_dst_aver_v_src1_16x16_msa;
+    c->avg_qpel_pixels_tab[0][15] =
+        ff_hv_mc_qpel_avg_dst_aver_hv_src11_16x16_msa;
+
+    c->avg_qpel_pixels_tab[1][0] = ff_avg_width8_msa;
+    c->avg_qpel_pixels_tab[1][1] =
+        ff_horiz_mc_qpel_avg_dst_aver_src0_8width_msa;
+    c->avg_qpel_pixels_tab[1][2] = ff_horiz_mc_qpel_avg_dst_8width_msa;
+    c->avg_qpel_pixels_tab[1][3] =
+        ff_horiz_mc_qpel_avg_dst_aver_src1_8width_msa;
+    c->avg_qpel_pixels_tab[1][4] = ff_vert_mc_qpel_avg_dst_aver_src0_8x8_msa;
+    c->avg_qpel_pixels_tab[1][5] = ff_hv_mc_qpel_avg_dst_aver_hv_src00_8x8_msa;
+    c->avg_qpel_pixels_tab[1][6] = ff_hv_mc_qpel_avg_dst_aver_v_src0_8x8_msa;
+    c->avg_qpel_pixels_tab[1][7] = ff_hv_mc_qpel_avg_dst_aver_hv_src10_8x8_msa;
+    c->avg_qpel_pixels_tab[1][8] = ff_vert_mc_qpel_avg_dst_8x8_msa;
+    c->avg_qpel_pixels_tab[1][9] = ff_hv_mc_qpel_avg_dst_aver_h_src0_8x8_msa;
+    c->avg_qpel_pixels_tab[1][10] = ff_hv_mc_qpel_avg_dst_8x8_msa;
+    c->avg_qpel_pixels_tab[1][11] = ff_hv_mc_qpel_avg_dst_aver_h_src1_8x8_msa;
+    c->avg_qpel_pixels_tab[1][12] = ff_vert_mc_qpel_avg_dst_aver_src1_8x8_msa;
+    c->avg_qpel_pixels_tab[1][13] = ff_hv_mc_qpel_avg_dst_aver_hv_src01_8x8_msa;
+    c->avg_qpel_pixels_tab[1][14] = ff_hv_mc_qpel_avg_dst_aver_v_src1_8x8_msa;
+    c->avg_qpel_pixels_tab[1][15] = ff_hv_mc_qpel_avg_dst_aver_hv_src11_8x8_msa;
+}
+#endif  // #if HAVE_MSA
+
+void ff_qpeldsp_init_mips(QpelDSPContext *c)
+{
+#if HAVE_MSA
+    qpeldsp_init_msa(c);
+#endif  // #if HAVE_MSA
+}

libavcodec/mips/qpeldsp_mips.h

@@ -0,0 +1,261 @@
+/*
+ * Copyright (c) 2015 Parag Salasakar (Parag.Salasakar@imgtec.com)
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVCODEC_MIPS_QPELDSP_MIPS_H
+#define AVCODEC_MIPS_QPELDSP_MIPS_H
+
+#include "../mpegvideo.h"
+
+void ff_copy_8x8_msa(uint8_t *dst, const uint8_t *src, ptrdiff_t stride);
+void ff_copy_16x16_msa(uint8_t *dst, const uint8_t *src, ptrdiff_t stride);
+void ff_avg_width8_msa(uint8_t *dst, const uint8_t *src, ptrdiff_t stride);
+void ff_avg_width16_msa(uint8_t *dst, const uint8_t *src, ptrdiff_t stride);
+
+void ff_horiz_mc_qpel_aver_src0_8width_msa(uint8_t *dst, const uint8_t *src,
+                                           ptrdiff_t stride);
+void ff_horiz_mc_qpel_aver_src0_16width_msa(uint8_t *dst, const uint8_t *src,
+                                            ptrdiff_t stride);
+void ff_horiz_mc_qpel_8width_msa(uint8_t *dst, const uint8_t *src,
+                                 ptrdiff_t stride);
+void ff_horiz_mc_qpel_16width_msa(uint8_t *dst, const uint8_t *src,
+                                  ptrdiff_t stride);
+void ff_horiz_mc_qpel_aver_src1_8width_msa(uint8_t *dst, const uint8_t *src,
+                                           ptrdiff_t stride);
+void ff_horiz_mc_qpel_aver_src1_16width_msa(uint8_t *dst, const uint8_t *src,
+                                            ptrdiff_t stride);
+void ff_horiz_mc_qpel_no_rnd_aver_src0_8width_msa(uint8_t *dst,
+                                                  const uint8_t *src,
+                                                  ptrdiff_t stride);
+void ff_horiz_mc_qpel_no_rnd_aver_src0_16width_msa(uint8_t *dst,
+                                                   const uint8_t *src,
+                                                   ptrdiff_t stride);
+void ff_horiz_mc_qpel_no_rnd_8width_msa(uint8_t *dst, const uint8_t *src,
+                                        ptrdiff_t stride);
+void ff_horiz_mc_qpel_no_rnd_16width_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_horiz_mc_qpel_no_rnd_aver_src1_8width_msa(uint8_t *dst,
+                                                  const uint8_t *src,
+                                                  ptrdiff_t stride);
+void ff_horiz_mc_qpel_no_rnd_aver_src1_16width_msa(uint8_t *dst,
+                                                   const uint8_t *src,
+                                                   ptrdiff_t stride);
+void ff_horiz_mc_qpel_avg_dst_aver_src0_8width_msa(uint8_t *dst,
+                                                   const uint8_t *src,
+                                                   ptrdiff_t stride);
+void ff_horiz_mc_qpel_avg_dst_aver_src0_16width_msa(uint8_t *dst,
+                                                    const uint8_t *src,
+                                                    ptrdiff_t stride);
+void ff_horiz_mc_qpel_avg_dst_8width_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_horiz_mc_qpel_avg_dst_16width_msa(uint8_t *dst, const uint8_t *src,
+                                          ptrdiff_t stride);
+void ff_horiz_mc_qpel_avg_dst_aver_src1_8width_msa(uint8_t *dst,
+                                                   const uint8_t *src,
+                                                   ptrdiff_t stride);
+void ff_horiz_mc_qpel_avg_dst_aver_src1_16width_msa(uint8_t *dst,
+                                                    const uint8_t *src,
+                                                    ptrdiff_t stride);
+
+void ff_vert_mc_qpel_aver_src0_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                       ptrdiff_t stride);
+void ff_vert_mc_qpel_aver_src0_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_vert_mc_qpel_8x8_msa(uint8_t *dst, const uint8_t *src,
+                             ptrdiff_t stride);
+void ff_vert_mc_qpel_16x16_msa(uint8_t *dst, const uint8_t *src,
+                               ptrdiff_t stride);
+void ff_vert_mc_qpel_aver_src1_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                       ptrdiff_t stride);
+void ff_vert_mc_qpel_aver_src1_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_vert_mc_qpel_no_rnd_aver_src0_8x8_msa(uint8_t *dst,
+                                              const uint8_t *src,
+                                              ptrdiff_t stride);
+void ff_vert_mc_qpel_no_rnd_aver_src0_16x16_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_vert_mc_qpel_no_rnd_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                    ptrdiff_t stride);
+void ff_vert_mc_qpel_no_rnd_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                      ptrdiff_t stride);
+void ff_vert_mc_qpel_no_rnd_aver_src1_8x8_msa(uint8_t *dst,
+                                              const uint8_t *src,
+                                              ptrdiff_t stride);
+void ff_vert_mc_qpel_no_rnd_aver_src1_16x16_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_vert_mc_qpel_avg_dst_aver_src0_8x8_msa(uint8_t *dst,
+                                               const uint8_t *src,
+                                               ptrdiff_t stride);
+void ff_vert_mc_qpel_avg_dst_aver_src0_16x16_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_vert_mc_qpel_avg_dst_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                     ptrdiff_t stride);
+void ff_vert_mc_qpel_avg_dst_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                       ptrdiff_t stride);
+void ff_vert_mc_qpel_avg_dst_aver_src1_8x8_msa(uint8_t *dst,
+                                               const uint8_t *src,
+                                               ptrdiff_t stride);
+void ff_vert_mc_qpel_avg_dst_aver_src1_16x16_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+
+void ff_hv_mc_qpel_aver_hv_src00_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                           ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_hv_src00_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_v_src0_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_v_src0_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                       ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_hv_src10_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                           ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_hv_src10_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_h_src0_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_h_src0_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                       ptrdiff_t stride);
+void ff_hv_mc_qpel_16x16_msa(uint8_t *dst, const uint8_t *src,
+                             ptrdiff_t stride);
+void ff_hv_mc_qpel_8x8_msa(uint8_t *dst, const uint8_t *src, ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_h_src1_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_h_src1_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                       ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_hv_src01_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                           ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_hv_src01_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_v_src1_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_v_src1_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                       ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_hv_src11_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                           ptrdiff_t stride);
+void ff_hv_mc_qpel_aver_hv_src11_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                         ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_hv_src00_16x16_msa(uint8_t *dst,
+                                                   const uint8_t *src,
+                                                   ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_hv_src00_8x8_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_v_src0_16x16_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_v_src0_8x8_msa(uint8_t *dst,
+                                               const uint8_t *src,
+                                               ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_hv_src10_16x16_msa(uint8_t *dst,
+                                                   const uint8_t *src,
+                                                   ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_hv_src10_8x8_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_h_src0_16x16_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_h_src0_8x8_msa(uint8_t *dst,
+                                               const uint8_t *src,
+                                               ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                     ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                   ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_h_src1_16x16_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_h_src1_8x8_msa(uint8_t *dst,
+                                               const uint8_t *src,
+                                               ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_hv_src01_16x16_msa(uint8_t *dst,
+                                                   const uint8_t *src,
+                                                   ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_hv_src01_8x8_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_v_src1_16x16_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_v_src1_8x8_msa(uint8_t *dst,
+                                               const uint8_t *src,
+                                               ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_hv_src11_16x16_msa(uint8_t *dst,
+                                                   const uint8_t *src,
+                                                   ptrdiff_t stride);
+void ff_hv_mc_qpel_avg_dst_aver_hv_src11_8x8_msa(uint8_t *dst,
+                                                 const uint8_t *src,
+                                                 ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_hv_src00_16x16_msa(uint8_t *dst,
+                                                  const uint8_t *src,
+                                                  ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_hv_src00_8x8_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_v_src0_16x16_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_v_src0_8x8_msa(uint8_t *dst,
+                                              const uint8_t *src,
+                                              ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_hv_src10_16x16_msa(uint8_t *dst,
+                                                  const uint8_t *src,
+                                                  ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_hv_src10_8x8_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_h_src0_16x16_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_h_src0_8x8_msa(uint8_t *dst,
+                                              const uint8_t *src,
+                                              ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_16x16_msa(uint8_t *dst, const uint8_t *src,
+                                    ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_8x8_msa(uint8_t *dst, const uint8_t *src,
+                                  ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_h_src1_16x16_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_h_src1_8x8_msa(uint8_t *dst,
+                                              const uint8_t *src,
+                                              ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_hv_src01_16x16_msa(uint8_t *dst,
+                                                  const uint8_t *src,
+                                                  ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_hv_src01_8x8_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_v_src1_16x16_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_v_src1_8x8_msa(uint8_t *dst,
+                                              const uint8_t *src,
+                                              ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_hv_src11_16x16_msa(uint8_t *dst,
+                                                  const uint8_t *src,
+                                                  ptrdiff_t stride);
+void ff_hv_mc_qpel_no_rnd_aver_hv_src11_8x8_msa(uint8_t *dst,
+                                                const uint8_t *src,
+                                                ptrdiff_t stride);
+
+#endif  // #ifndef AVCODEC_MIPS_QPELDSP_MIPS_H

libavcodec/qpeldsp.c

@@ -811,4 +811,6 @@ av_cold void ff_qpeldsp_init(QpelDSPContext *c)
 
     if (ARCH_X86)
         ff_qpeldsp_init_x86(c);
+    if (ARCH_MIPS)
+        ff_qpeldsp_init_mips(c);
 }

libavcodec/qpeldsp.h

@@ -78,5 +78,6 @@ typedef struct QpelDSPContext {
 void ff_qpeldsp_init(QpelDSPContext *c);
 
 void ff_qpeldsp_init_x86(QpelDSPContext *c);
+void ff_qpeldsp_init_mips(QpelDSPContext *c);
 
 #endif /* AVCODEC_QPELDSP_H */

libavutil/mips/generic_macros_msa.h

@@ -1027,6 +1027,27 @@
 }
 #define DPADD_SB4_SH(...) DPADD_SB4(v8i16, __VA_ARGS__)
 
+/* Description : Dot product & addition of byte vector elements
+   Arguments   : Inputs  - mult0, mult1
+                           cnst0, cnst1
+                 Outputs - out0, out1
+                 Return Type - unsigned halfword
+   Details     : Unsigned byte elements from mult0 are multiplied with
+                 unsigned byte elements from cnst0 producing a result
+                 twice the size of input i.e. unsigned halfword.
+                 Then this multiplication results of adjacent odd-even elements
+                 are added to the out vector
+                 (2 unsigned halfword results)
+*/
+#define DPADD_UB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)   \
+{                                                                  \
+    out0 = (RTYPE) __msa_dpadd_u_h((v8u16) out0,                   \
+                                   (v16u8) mult0, (v16u8) cnst0);  \
+    out1 = (RTYPE) __msa_dpadd_u_h((v8u16) out1,                   \
+                                   (v16u8) mult1, (v16u8) cnst1);  \
+}
+#define DPADD_UB2_UH(...) DPADD_UB2(v8u16, __VA_ARGS__)
+
 /* Description : Dot product & addition of halfword vector elements
    Arguments   : Inputs  - mult0, mult1
                            cnst0, cnst1