diff options
Diffstat (limited to 'vp9')
-rw-r--r-- | vp9/common/arm/neon/vp9_save_reg_neon.asm | 36 | ||||
-rw-r--r-- | vp9/common/mips/dspr2/vp9_itrans4_dspr2.c | 8 | ||||
-rw-r--r-- | vp9/common/mips/msa/vp9_idct16x16_msa.c | 16 | ||||
-rw-r--r-- | vp9/common/vp9_idct.c | 76 | ||||
-rw-r--r-- | vp9/common/vp9_idct.h | 4 | ||||
-rw-r--r-- | vp9/vp9_common.mk | 2 |
6 files changed, 52 insertions, 90 deletions
diff --git a/vp9/common/arm/neon/vp9_save_reg_neon.asm b/vp9/common/arm/neon/vp9_save_reg_neon.asm deleted file mode 100644 index 71c3e7077..000000000 --- a/vp9/common/arm/neon/vp9_save_reg_neon.asm +++ /dev/null @@ -1,36 +0,0 @@ -; -; Copyright (c) 2010 The WebM project authors. All Rights Reserved. -; -; Use of this source code is governed by a BSD-style license -; that can be found in the LICENSE file in the root of the source -; tree. An additional intellectual property rights grant can be found -; in the file PATENTS. All contributing project authors may -; be found in the AUTHORS file in the root of the source tree. -; - - - EXPORT |vp9_push_neon| - EXPORT |vp9_pop_neon| - - ARM - REQUIRE8 - PRESERVE8 - - AREA ||.text||, CODE, READONLY, ALIGN=2 - -|vp9_push_neon| PROC - vst1.i64 {d8, d9, d10, d11}, [r0]! - vst1.i64 {d12, d13, d14, d15}, [r0]! - bx lr - - ENDP - -|vp9_pop_neon| PROC - vld1.i64 {d8, d9, d10, d11}, [r0]! - vld1.i64 {d12, d13, d14, d15}, [r0]! - bx lr - - ENDP - - END - diff --git a/vp9/common/mips/dspr2/vp9_itrans4_dspr2.c b/vp9/common/mips/dspr2/vp9_itrans4_dspr2.c index 848f7c0aa..c10979b64 100644 --- a/vp9/common/mips/dspr2/vp9_itrans4_dspr2.c +++ b/vp9/common/mips/dspr2/vp9_itrans4_dspr2.c @@ -38,11 +38,11 @@ void vp9_iht4x4_16_add_dspr2(const int16_t *input, uint8_t *dest, switch (tx_type) { case DCT_DCT: // DCT in both horizontal and vertical - vp9_idct4_rows_dspr2(input, outptr); - vp9_idct4_columns_add_blk_dspr2(&out[0], dest, dest_stride); + vpx_idct4_rows_dspr2(input, outptr); + vpx_idct4_columns_add_blk_dspr2(&out[0], dest, dest_stride); break; case ADST_DCT: // ADST in vertical, DCT in horizontal - vp9_idct4_rows_dspr2(input, outptr); + vpx_idct4_rows_dspr2(input, outptr); outptr = out; @@ -69,7 +69,7 @@ void vp9_iht4x4_16_add_dspr2(const int16_t *input, uint8_t *dest, temp_in[i * 4 + j] = out[j * 4 + i]; } } - vp9_idct4_columns_add_blk_dspr2(&temp_in[0], dest, dest_stride); + vpx_idct4_columns_add_blk_dspr2(&temp_in[0], dest, dest_stride); break; case ADST_ADST: // ADST in both directions for (i = 0; i < 4; ++i) { diff --git a/vp9/common/mips/msa/vp9_idct16x16_msa.c b/vp9/common/mips/msa/vp9_idct16x16_msa.c index 59b8930b5..5adf0aaac 100644 --- a/vp9/common/mips/msa/vp9_idct16x16_msa.c +++ b/vp9/common/mips/msa/vp9_idct16x16_msa.c @@ -24,13 +24,13 @@ void vp9_iht16x16_256_add_msa(const int16_t *input, uint8_t *dst, /* transform rows */ for (i = 0; i < 2; ++i) { /* process 16 * 8 block */ - vp9_idct16_1d_rows_msa((input + (i << 7)), (out_ptr + (i << 7))); + vpx_idct16_1d_rows_msa((input + (i << 7)), (out_ptr + (i << 7))); } /* transform columns */ for (i = 0; i < 2; ++i) { /* process 8 * 16 block */ - vp9_idct16_1d_columns_addblk_msa((out_ptr + (i << 3)), (dst + (i << 3)), + vpx_idct16_1d_columns_addblk_msa((out_ptr + (i << 3)), (dst + (i << 3)), dst_stride); } break; @@ -38,12 +38,12 @@ void vp9_iht16x16_256_add_msa(const int16_t *input, uint8_t *dst, /* transform rows */ for (i = 0; i < 2; ++i) { /* process 16 * 8 block */ - vp9_idct16_1d_rows_msa((input + (i << 7)), (out_ptr + (i << 7))); + vpx_idct16_1d_rows_msa((input + (i << 7)), (out_ptr + (i << 7))); } /* transform columns */ for (i = 0; i < 2; ++i) { - vp9_iadst16_1d_columns_addblk_msa((out_ptr + (i << 3)), + vpx_iadst16_1d_columns_addblk_msa((out_ptr + (i << 3)), (dst + (i << 3)), dst_stride); } break; @@ -51,13 +51,13 @@ void vp9_iht16x16_256_add_msa(const int16_t *input, uint8_t *dst, /* transform rows */ for (i = 0; i < 2; ++i) { /* process 16 * 8 block */ - vp9_iadst16_1d_rows_msa((input + (i << 7)), (out_ptr + (i << 7))); + vpx_iadst16_1d_rows_msa((input + (i << 7)), (out_ptr + (i << 7))); } /* transform columns */ for (i = 0; i < 2; ++i) { /* process 8 * 16 block */ - vp9_idct16_1d_columns_addblk_msa((out_ptr + (i << 3)), (dst + (i << 3)), + vpx_idct16_1d_columns_addblk_msa((out_ptr + (i << 3)), (dst + (i << 3)), dst_stride); } break; @@ -65,12 +65,12 @@ void vp9_iht16x16_256_add_msa(const int16_t *input, uint8_t *dst, /* transform rows */ for (i = 0; i < 2; ++i) { /* process 16 * 8 block */ - vp9_iadst16_1d_rows_msa((input + (i << 7)), (out_ptr + (i << 7))); + vpx_iadst16_1d_rows_msa((input + (i << 7)), (out_ptr + (i << 7))); } /* transform columns */ for (i = 0; i < 2; ++i) { - vp9_iadst16_1d_columns_addblk_msa((out_ptr + (i << 3)), + vpx_iadst16_1d_columns_addblk_msa((out_ptr + (i << 3)), (dst + (i << 3)), dst_stride); } break; diff --git a/vp9/common/vp9_idct.c b/vp9/common/vp9_idct.c index b15f7f370..146bf4298 100644 --- a/vp9/common/vp9_idct.c +++ b/vp9/common/vp9_idct.c @@ -123,18 +123,18 @@ void vp9_iht16x16_256_add_c(const tran_low_t *input, uint8_t *dest, int stride, void vp9_idct4x4_add(const tran_low_t *input, uint8_t *dest, int stride, int eob) { if (eob > 1) - vp9_idct4x4_16_add(input, dest, stride); + vpx_idct4x4_16_add(input, dest, stride); else - vp9_idct4x4_1_add(input, dest, stride); + vpx_idct4x4_1_add(input, dest, stride); } void vp9_iwht4x4_add(const tran_low_t *input, uint8_t *dest, int stride, int eob) { if (eob > 1) - vp9_iwht4x4_16_add(input, dest, stride); + vpx_iwht4x4_16_add(input, dest, stride); else - vp9_iwht4x4_1_add(input, dest, stride); + vpx_iwht4x4_1_add(input, dest, stride); } void vp9_idct8x8_add(const tran_low_t *input, uint8_t *dest, int stride, @@ -148,11 +148,11 @@ void vp9_idct8x8_add(const tran_low_t *input, uint8_t *dest, int stride, // Combine that with code here. if (eob == 1) // DC only DCT coefficient - vp9_idct8x8_1_add(input, dest, stride); + vpx_idct8x8_1_add(input, dest, stride); else if (eob <= 12) - vp9_idct8x8_12_add(input, dest, stride); + vpx_idct8x8_12_add(input, dest, stride); else - vp9_idct8x8_64_add(input, dest, stride); + vpx_idct8x8_64_add(input, dest, stride); } void vp9_idct16x16_add(const tran_low_t *input, uint8_t *dest, int stride, @@ -161,22 +161,22 @@ void vp9_idct16x16_add(const tran_low_t *input, uint8_t *dest, int stride, * coefficients. Use eobs to separate different cases. */ if (eob == 1) /* DC only DCT coefficient. */ - vp9_idct16x16_1_add(input, dest, stride); + vpx_idct16x16_1_add(input, dest, stride); else if (eob <= 10) - vp9_idct16x16_10_add(input, dest, stride); + vpx_idct16x16_10_add(input, dest, stride); else - vp9_idct16x16_256_add(input, dest, stride); + vpx_idct16x16_256_add(input, dest, stride); } void vp9_idct32x32_add(const tran_low_t *input, uint8_t *dest, int stride, int eob) { if (eob == 1) - vp9_idct32x32_1_add(input, dest, stride); + vpx_idct32x32_1_add(input, dest, stride); else if (eob <= 34) // non-zero coeff only in upper-left 8x8 - vp9_idct32x32_34_add(input, dest, stride); + vpx_idct32x32_34_add(input, dest, stride); else - vp9_idct32x32_1024_add(input, dest, stride); + vpx_idct32x32_1024_add(input, dest, stride); } // iht @@ -210,10 +210,10 @@ void vp9_iht16x16_add(TX_TYPE tx_type, const tran_low_t *input, uint8_t *dest, void vp9_highbd_iht4x4_16_add_c(const tran_low_t *input, uint8_t *dest8, int stride, int tx_type, int bd) { const highbd_transform_2d IHT_4[] = { - { vp9_highbd_idct4_c, vp9_highbd_idct4_c }, // DCT_DCT = 0 - { highbd_iadst4_c, vp9_highbd_idct4_c }, // ADST_DCT = 1 - { vp9_highbd_idct4_c, highbd_iadst4_c }, // DCT_ADST = 2 - { highbd_iadst4_c, highbd_iadst4_c } // ADST_ADST = 3 + { vpx_highbd_idct4_c, vpx_highbd_idct4_c }, // DCT_DCT = 0 + { vpx_highbd_iadst4_c, vpx_highbd_idct4_c }, // ADST_DCT = 1 + { vpx_highbd_idct4_c, vpx_highbd_iadst4_c }, // DCT_ADST = 2 + { vpx_highbd_iadst4_c, vpx_highbd_iadst4_c } // ADST_ADST = 3 }; uint16_t *dest = CONVERT_TO_SHORTPTR(dest8); @@ -242,10 +242,10 @@ void vp9_highbd_iht4x4_16_add_c(const tran_low_t *input, uint8_t *dest8, } static const highbd_transform_2d HIGH_IHT_8[] = { - { vp9_highbd_idct8_c, vp9_highbd_idct8_c }, // DCT_DCT = 0 - { highbd_iadst8_c, vp9_highbd_idct8_c }, // ADST_DCT = 1 - { vp9_highbd_idct8_c, highbd_iadst8_c }, // DCT_ADST = 2 - { highbd_iadst8_c, highbd_iadst8_c } // ADST_ADST = 3 + { vpx_highbd_idct8_c, vpx_highbd_idct8_c }, // DCT_DCT = 0 + { vpx_highbd_iadst8_c, vpx_highbd_idct8_c }, // ADST_DCT = 1 + { vpx_highbd_idct8_c, vpx_highbd_iadst8_c }, // DCT_ADST = 2 + { vpx_highbd_iadst8_c, vpx_highbd_iadst8_c } // ADST_ADST = 3 }; void vp9_highbd_iht8x8_64_add_c(const tran_low_t *input, uint8_t *dest8, @@ -277,10 +277,10 @@ void vp9_highbd_iht8x8_64_add_c(const tran_low_t *input, uint8_t *dest8, } static const highbd_transform_2d HIGH_IHT_16[] = { - { vp9_highbd_idct16_c, vp9_highbd_idct16_c }, // DCT_DCT = 0 - { highbd_iadst16_c, vp9_highbd_idct16_c }, // ADST_DCT = 1 - { vp9_highbd_idct16_c, highbd_iadst16_c }, // DCT_ADST = 2 - { highbd_iadst16_c, highbd_iadst16_c } // ADST_ADST = 3 + { vpx_highbd_idct16_c, vpx_highbd_idct16_c }, // DCT_DCT = 0 + { vpx_highbd_iadst16_c, vpx_highbd_idct16_c }, // ADST_DCT = 1 + { vpx_highbd_idct16_c, vpx_highbd_iadst16_c }, // DCT_ADST = 2 + { vpx_highbd_iadst16_c, vpx_highbd_iadst16_c } // ADST_ADST = 3 }; void vp9_highbd_iht16x16_256_add_c(const tran_low_t *input, uint8_t *dest8, @@ -315,18 +315,18 @@ void vp9_highbd_iht16x16_256_add_c(const tran_low_t *input, uint8_t *dest8, void vp9_highbd_idct4x4_add(const tran_low_t *input, uint8_t *dest, int stride, int eob, int bd) { if (eob > 1) - vp9_highbd_idct4x4_16_add(input, dest, stride, bd); + vpx_highbd_idct4x4_16_add(input, dest, stride, bd); else - vp9_highbd_idct4x4_1_add(input, dest, stride, bd); + vpx_highbd_idct4x4_1_add(input, dest, stride, bd); } void vp9_highbd_iwht4x4_add(const tran_low_t *input, uint8_t *dest, int stride, int eob, int bd) { if (eob > 1) - vp9_highbd_iwht4x4_16_add(input, dest, stride, bd); + vpx_highbd_iwht4x4_16_add(input, dest, stride, bd); else - vp9_highbd_iwht4x4_1_add(input, dest, stride, bd); + vpx_highbd_iwht4x4_1_add(input, dest, stride, bd); } void vp9_highbd_idct8x8_add(const tran_low_t *input, uint8_t *dest, int stride, @@ -340,11 +340,11 @@ void vp9_highbd_idct8x8_add(const tran_low_t *input, uint8_t *dest, int stride, // Combine that with code here. // DC only DCT coefficient if (eob == 1) { - vp9_highbd_idct8x8_1_add(input, dest, stride, bd); + vpx_highbd_idct8x8_1_add(input, dest, stride, bd); } else if (eob <= 10) { - vp9_highbd_idct8x8_10_add(input, dest, stride, bd); + vpx_highbd_idct8x8_10_add(input, dest, stride, bd); } else { - vp9_highbd_idct8x8_64_add(input, dest, stride, bd); + vpx_highbd_idct8x8_64_add(input, dest, stride, bd); } } @@ -354,11 +354,11 @@ void vp9_highbd_idct16x16_add(const tran_low_t *input, uint8_t *dest, // coefficients. Use eobs to separate different cases. // DC only DCT coefficient. if (eob == 1) { - vp9_highbd_idct16x16_1_add(input, dest, stride, bd); + vpx_highbd_idct16x16_1_add(input, dest, stride, bd); } else if (eob <= 10) { - vp9_highbd_idct16x16_10_add(input, dest, stride, bd); + vpx_highbd_idct16x16_10_add(input, dest, stride, bd); } else { - vp9_highbd_idct16x16_256_add(input, dest, stride, bd); + vpx_highbd_idct16x16_256_add(input, dest, stride, bd); } } @@ -366,11 +366,11 @@ void vp9_highbd_idct32x32_add(const tran_low_t *input, uint8_t *dest, int stride, int eob, int bd) { // Non-zero coeff only in upper-left 8x8 if (eob == 1) { - vp9_highbd_idct32x32_1_add(input, dest, stride, bd); + vpx_highbd_idct32x32_1_add(input, dest, stride, bd); } else if (eob <= 34) { - vp9_highbd_idct32x32_34_add(input, dest, stride, bd); + vpx_highbd_idct32x32_34_add(input, dest, stride, bd); } else { - vp9_highbd_idct32x32_1024_add(input, dest, stride, bd); + vpx_highbd_idct32x32_1024_add(input, dest, stride, bd); } } diff --git a/vp9/common/vp9_idct.h b/vp9/common/vp9_idct.h index 7a7dc1d64..b5a3fbf36 100644 --- a/vp9/common/vp9_idct.h +++ b/vp9/common/vp9_idct.h @@ -44,8 +44,8 @@ void vp9_idct4x4_add(const tran_low_t *input, uint8_t *dest, int stride, int eob); void vp9_idct8x8_add(const tran_low_t *input, uint8_t *dest, int stride, int eob); -void vp9_idct16x16_add(const tran_low_t *input, uint8_t *dest, int stride, int - eob); +void vp9_idct16x16_add(const tran_low_t *input, uint8_t *dest, int stride, + int eob); void vp9_idct32x32_add(const tran_low_t *input, uint8_t *dest, int stride, int eob); diff --git a/vp9/vp9_common.mk b/vp9/vp9_common.mk index 735aaf141..a49cd0aee 100644 --- a/vp9/vp9_common.mk +++ b/vp9/vp9_common.mk @@ -88,8 +88,6 @@ endif VP9_COMMON_SRCS-$(HAVE_SSE2) += common/x86/vp9_idct_intrin_sse2.c -VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_save_reg_neon$(ASM) - ifneq ($(CONFIG_VP9_HIGHBITDEPTH),yes) VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_iht4x4_add_neon.c VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_iht8x8_add_neon.c |