Quelle  intrapred4_dspr2.c   Sprache: C

/*
 *  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 "vpx_dsp/mips/common_dspr2.h"

#if HAVE_DSPR2
void vpx_h_predictor_4x4_dspr2(uint8_t *dst, ptrdiff_t stride,
                               const uint8_t *above, const uint8_t *left) {
  int32_t tmp1, tmp2, tmp3, tmp4;
  (void)above;

  __asm__ __volatile__(
      "lb %[tmp1], (%[left]) \n\t"
      "lb %[tmp2], 1(%[left]) \n\t"
      "lb %[tmp3], 2(%[left]) \n\t"
      "lb %[tmp4], 3(%[left]) \n\t"
      "replv.qb %[tmp1], %[tmp1] \n\t"
      "replv.qb %[tmp2], %[tmp2] \n\t"
      "replv.qb %[tmp3], %[tmp3] \n\t"
      "replv.qb %[tmp4], %[tmp4] \n\t"
      "sw %[tmp1], (%[dst]) \n\t"
      "add %[dst], %[dst], %[stride] \n\t"
      "sw %[tmp2], (%[dst]) \n\t"
      "add %[dst], %[dst], %[stride] \n\t"
      "sw %[tmp3], (%[dst]) \n\t"
      "add %[dst], %[dst], %[stride] \n\t"
      "sw %[tmp4], (%[dst]) \n\t"

      : [tmp1] "=&r"(tmp1), [tmp2] "=&r"(tmp2), [tmp3] "=&r"(tmp3),
        [tmp4] "=&r"(tmp4)
      : [left] "r"(left), [dst] "r"(dst), [stride] "r"(stride));
}

void vpx_dc_predictor_4x4_dspr2(uint8_t *dst, ptrdiff_t stride,
                                const uint8_t *above, const uint8_t *left) {
  int32_t expected_dc;
  int32_t average;
  int32_t tmp, above_c, above_l, above_r, left_c, left_r, left_l;

  __asm__ __volatile__(
      "lw %[above_c], (%[above]) \n\t"
      "lw %[left_c], (%[left]) \n\t"

      "preceu.ph.qbl %[above_l], %[above_c] \n\t"
      "preceu.ph.qbr %[above_r], %[above_c] \n\t"
      "preceu.ph.qbl %[left_l], %[left_c] \n\t"
      "preceu.ph.qbr %[left_r], %[left_c] \n\t"

      "addu.ph %[average], %[above_r], %[above_l] \n\t"
      "addu.ph %[average], %[average], %[left_l] \n\t"
      "addu.ph %[average], %[average], %[left_r] \n\t"
      "addiu %[average], %[average], 4 \n\t"
      "srl %[tmp], %[average], 16 \n\t"
      "addu.ph %[average], %[tmp], %[average] \n\t"
      "srl %[expected_dc], %[average], 3 \n\t"
      "replv.qb %[expected_dc], %[expected_dc] \n\t"

      "sw %[expected_dc], (%[dst]) \n\t"
      "add %[dst], %[dst], %[stride] \n\t"
      "sw %[expected_dc], (%[dst]) \n\t"
      "add %[dst], %[dst], %[stride] \n\t"
      "sw %[expected_dc], (%[dst]) \n\t"
      "add %[dst], %[dst], %[stride] \n\t"
      "sw %[expected_dc], (%[dst]) \n\t"

      : [above_c] "=&r"(above_c), [above_l] "=&r"(above_l),
        [above_r] "=&r"(above_r), [left_c] "=&r"(left_c),
        [left_l] "=&r"(left_l), [left_r] "=&r"(left_r),
        [average] "=&r"(average), [tmp] "=&r"(tmp),
        [expected_dc] "=&r"(expected_dc)
      : [above] "r"(above), [left] "r"(left), [dst] "r"(dst),
        [stride] "r"(stride));
}

void vpx_tm_predictor_4x4_dspr2(uint8_t *dst, ptrdiff_t stride,
                                const uint8_t *above, const uint8_t *left) {
  int32_t abovel, abover;
  int32_t left0, left1, left2, left3;
  int32_t res0, res1;
  int32_t resl;
  int32_t resr;
  int32_t top_left;
  uint8_t *cm = vpx_ff_cropTbl;

  __asm__ __volatile__(
      "ulw %[resl], (%[above]) \n\t"

      "lbu %[left0], (%[left]) \n\t"
      "lbu %[left1], 1(%[left]) \n\t"
      "lbu %[left2], 2(%[left]) \n\t"
      "lbu %[left3], 3(%[left]) \n\t"

      "lbu %[top_left], -1(%[above]) \n\t"

      "preceu.ph.qbl %[abovel], %[resl] \n\t"
      "preceu.ph.qbr %[abover], %[resl] \n\t"

      "replv.ph %[left0], %[left0] \n\t"
      "replv.ph %[left1], %[left1] \n\t"
      "replv.ph %[left2], %[left2] \n\t"
      "replv.ph %[left3], %[left3] \n\t"

      "replv.ph %[top_left], %[top_left] \n\t"

      "addu.ph %[resl], %[abovel], %[left0] \n\t"
      "subu.ph %[resl], %[resl], %[top_left] \n\t"

      "addu.ph %[resr], %[abover], %[left0] \n\t"
      "subu.ph %[resr], %[resr], %[top_left] \n\t"

      "sll %[res0], %[resr], 16 \n\t"
      "sra %[res0], %[res0], 16 \n\t"
      "lbux %[res0], %[res0](%[cm]) \n\t"

      "sra %[res1], %[resr], 16 \n\t"
      "lbux %[res1], %[res1](%[cm]) \n\t"
      "sb %[res0], (%[dst]) \n\t"

      "sll %[res0], %[resl], 16 \n\t"
      "sra %[res0], %[res0], 16 \n\t"
      "lbux %[res0], %[res0](%[cm]) \n\t"
      "sb %[res1], 1(%[dst]) \n\t"

      "sra %[res1], %[resl], 16 \n\t"
      "lbux %[res1], %[res1](%[cm]) \n\t"

      "addu.ph %[resl], %[abovel], %[left1] \n\t"
      "subu.ph %[resl], %[resl], %[top_left] \n\t"

      "addu.ph %[resr], %[abover], %[left1] \n\t"
      "subu.ph %[resr], %[resr], %[top_left] \n\t"

      "sb %[res0], 2(%[dst]) \n\t"
      "sb %[res1], 3(%[dst]) \n\t"

      "add %[dst], %[dst], %[stride] \n\t"

      "sll %[res0], %[resr], 16 \n\t"
      "sra %[res0], %[res0], 16 \n\t"
      "lbux %[res0], %[res0](%[cm]) \n\t"

      "sra %[res1], %[resr], 16 \n\t"
      "lbux %[res1], %[res1](%[cm]) \n\t"
      "sb %[res0], (%[dst]) \n\t"

      "sll %[res0], %[resl], 16 \n\t"
      "sra %[res0], %[res0], 16 \n\t"
      "lbux %[res0], %[res0](%[cm]) \n\t"

      "sb %[res1], 1(%[dst]) \n\t"
      "sra %[res1], %[resl], 16 \n\t"
      "lbux %[res1], %[res1](%[cm]) \n\t"

      "addu.ph %[resl], %[abovel], %[left2] \n\t"
      "subu.ph %[resl], %[resl], %[top_left] \n\t"

      "addu.ph %[resr], %[abover], %[left2] \n\t"
      "subu.ph %[resr], %[resr], %[top_left] \n\t"

      "sb %[res0], 2(%[dst]) \n\t"
      "sb %[res1], 3(%[dst]) \n\t"

      "add %[dst], %[dst], %[stride] \n\t"

      "sll %[res0], %[resr], 16 \n\t"
      "sra %[res0], %[res0], 16 \n\t"
      "lbux %[res0], %[res0](%[cm]) \n\t"

      "sra %[res1], %[resr], 16 \n\t"
      "lbux %[res1], %[res1](%[cm]) \n\t"
      "sb %[res0], (%[dst]) \n\t"

      "sll %[res0], %[resl], 16 \n\t"
      "sra %[res0], %[res0], 16 \n\t"
      "lbux %[res0], %[res0](%[cm]) \n\t"

      "sb %[res1], 1(%[dst]) \n\t"
      "sra %[res1], %[resl], 16 \n\t"
      "lbux %[res1], %[res1](%[cm]) \n\t"

      "addu.ph %[resl], %[abovel], %[left3] \n\t"
      "subu.ph %[resl], %[resl], %[top_left] \n\t"

      "addu.ph %[resr], %[abover], %[left3] \n\t"
      "subu.ph %[resr], %[resr], %[top_left] \n\t"

      "sb %[res0], 2(%[dst]) \n\t"
      "sb %[res1], 3(%[dst]) \n\t"

      "add %[dst], %[dst], %[stride] \n\t"

      "sll %[res0], %[resr], 16 \n\t"
      "sra %[res0], %[res0], 16 \n\t"
      "lbux %[res0], %[res0](%[cm]) \n\t"

      "sra %[res1], %[resr], 16 \n\t"
      "lbux %[res1], %[res1](%[cm]) \n\t"
      "sb %[res0], (%[dst]) \n\t"

      "sll %[res0], %[resl], 16 \n\t"
      "sra %[res0], %[res0], 16 \n\t"
      "lbux %[res0], %[res0](%[cm]) \n\t"
      "sb %[res1], 1(%[dst]) \n\t"

      "sra %[res1], %[resl], 16 \n\t"
      "lbux %[res1], %[res1](%[cm]) \n\t"

      "sb %[res0], 2(%[dst]) \n\t"
      "sb %[res1], 3(%[dst]) \n\t"

      : [abovel] "=&r"(abovel), [abover] "=&r"(abover), [left0] "=&r"(left0),
        [left1] "=&r"(left1), [left2] "=&r"(left2), [res0] "=&r"(res0),
        [res1] "=&r"(res1), [left3] "=&r"(left3), [resl] "=&r"(resl),
        [resr] "=&r"(resr), [top_left] "=&r"(top_left)
      : [above] "r"(above), [left] "r"(left), [dst] "r"(dst),
        [stride] "r"(stride), [cm] "r"(cm));
}
#endif  // #if HAVE_DSPR2