diff options
Diffstat (limited to 'config/arm-neon/vpx_dsp/arm/intrapred_neon_asm.asm.S')
| -rw-r--r-- | config/arm-neon/vpx_dsp/arm/intrapred_neon_asm.asm.S | 124 |
1 files changed, 56 insertions, 68 deletions
diff --git a/config/arm-neon/vpx_dsp/arm/intrapred_neon_asm.asm.S b/config/arm-neon/vpx_dsp/arm/intrapred_neon_asm.asm.S index 9f656b803..130dcd936 100644 --- a/config/arm-neon/vpx_dsp/arm/intrapred_neon_asm.asm.S +++ b/config/arm-neon/vpx_dsp/arm/intrapred_neon_asm.asm.S @@ -1,6 +1,6 @@ @ This file was created from a .asm file @ using the ads2gas.pl script. - .syntax unified +.syntax unified @ @ Copyright (c) 2014 The WebM project authors. All Rights Reserved. @ @@ -11,36 +11,36 @@ @ be found in the AUTHORS file in the root of the source tree. @ - .global vpx_v_predictor_4x4_neon - .type vpx_v_predictor_4x4_neon, function - .global vpx_v_predictor_8x8_neon - .type vpx_v_predictor_8x8_neon, function - .global vpx_v_predictor_16x16_neon - .type vpx_v_predictor_16x16_neon, function - .global vpx_v_predictor_32x32_neon - .type vpx_v_predictor_32x32_neon, function - .global vpx_h_predictor_4x4_neon - .type vpx_h_predictor_4x4_neon, function - .global vpx_h_predictor_8x8_neon - .type vpx_h_predictor_8x8_neon, function - .global vpx_h_predictor_16x16_neon - .type vpx_h_predictor_16x16_neon, function - .global vpx_h_predictor_32x32_neon - .type vpx_h_predictor_32x32_neon, function - .global vpx_tm_predictor_4x4_neon - .type vpx_tm_predictor_4x4_neon, function - .global vpx_tm_predictor_8x8_neon - .type vpx_tm_predictor_8x8_neon, function - .global vpx_tm_predictor_16x16_neon - .type vpx_tm_predictor_16x16_neon, function - .global vpx_tm_predictor_32x32_neon - .type vpx_tm_predictor_32x32_neon, function - .arm - .eabi_attribute 24, 1 @Tag_ABI_align_needed - .eabi_attribute 25, 1 @Tag_ABI_align_preserved - -.text -.p2align 2 + .global vpx_v_predictor_4x4_neon + .type vpx_v_predictor_4x4_neon, function + .global vpx_v_predictor_8x8_neon + .type vpx_v_predictor_8x8_neon, function + .global vpx_v_predictor_16x16_neon + .type vpx_v_predictor_16x16_neon, function + .global vpx_v_predictor_32x32_neon + .type vpx_v_predictor_32x32_neon, function + .global vpx_h_predictor_4x4_neon + .type vpx_h_predictor_4x4_neon, function + .global vpx_h_predictor_8x8_neon + .type vpx_h_predictor_8x8_neon, function + .global vpx_h_predictor_16x16_neon + .type vpx_h_predictor_16x16_neon, function + .global vpx_h_predictor_32x32_neon + .type vpx_h_predictor_32x32_neon, function + .global vpx_tm_predictor_4x4_neon + .type vpx_tm_predictor_4x4_neon, function + .global vpx_tm_predictor_8x8_neon + .type vpx_tm_predictor_8x8_neon, function + .global vpx_tm_predictor_16x16_neon + .type vpx_tm_predictor_16x16_neon, function + .global vpx_tm_predictor_32x32_neon + .type vpx_tm_predictor_32x32_neon, function + .arm + .eabi_attribute 24, 1 @Tag_ABI_align_needed + .eabi_attribute 25, 1 @Tag_ABI_align_preserved + + .text + .p2align 2 @void vpx_v_predictor_4x4_neon(uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -50,15 +50,14 @@ @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_v_predictor_4x4_neon: - vpx_v_predictor_4x4_neon: @ PROC +vpx_v_predictor_4x4_neon: @ PROC vld1.32 {d0[0]}, [r2] vst1.32 {d0[0]}, [r0], r1 vst1.32 {d0[0]}, [r0], r1 vst1.32 {d0[0]}, [r0], r1 vst1.32 {d0[0]}, [r0], r1 bx lr - .size vpx_v_predictor_4x4_neon, .-vpx_v_predictor_4x4_neon @ ENDP @ |vpx_v_predictor_4x4_neon| +.size vpx_v_predictor_4x4_neon, .-vpx_v_predictor_4x4_neon @ ENDP @ |vpx_v_predictor_4x4_neon| @void vpx_v_predictor_8x8_neon(uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -68,8 +67,7 @@ _vpx_v_predictor_4x4_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_v_predictor_8x8_neon: - vpx_v_predictor_8x8_neon: @ PROC +vpx_v_predictor_8x8_neon: @ PROC vld1.8 {d0}, [r2] vst1.8 {d0}, [r0], r1 vst1.8 {d0}, [r0], r1 @@ -80,7 +78,7 @@ _vpx_v_predictor_8x8_neon: vst1.8 {d0}, [r0], r1 vst1.8 {d0}, [r0], r1 bx lr - .size vpx_v_predictor_8x8_neon, .-vpx_v_predictor_8x8_neon @ ENDP @ |vpx_v_predictor_8x8_neon| +.size vpx_v_predictor_8x8_neon, .-vpx_v_predictor_8x8_neon @ ENDP @ |vpx_v_predictor_8x8_neon| @void vpx_v_predictor_16x16_neon(uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -90,8 +88,7 @@ _vpx_v_predictor_8x8_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_v_predictor_16x16_neon: - vpx_v_predictor_16x16_neon: @ PROC +vpx_v_predictor_16x16_neon: @ PROC vld1.8 {q0}, [r2] vst1.8 {q0}, [r0], r1 vst1.8 {q0}, [r0], r1 @@ -110,7 +107,7 @@ _vpx_v_predictor_16x16_neon: vst1.8 {q0}, [r0], r1 vst1.8 {q0}, [r0], r1 bx lr - .size vpx_v_predictor_16x16_neon, .-vpx_v_predictor_16x16_neon @ ENDP @ |vpx_v_predictor_16x16_neon| +.size vpx_v_predictor_16x16_neon, .-vpx_v_predictor_16x16_neon @ ENDP @ |vpx_v_predictor_16x16_neon| @void vpx_v_predictor_32x32_neon(uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -120,8 +117,7 @@ _vpx_v_predictor_16x16_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_v_predictor_32x32_neon: - vpx_v_predictor_32x32_neon: @ PROC +vpx_v_predictor_32x32_neon: @ PROC vld1.8 {q0, q1}, [r2] mov r2, #2 loop_v: @@ -144,7 +140,7 @@ loop_v: subs r2, r2, #1 bgt loop_v bx lr - .size vpx_v_predictor_32x32_neon, .-vpx_v_predictor_32x32_neon @ ENDP @ |vpx_v_predictor_32x32_neon| +.size vpx_v_predictor_32x32_neon, .-vpx_v_predictor_32x32_neon @ ENDP @ |vpx_v_predictor_32x32_neon| @void vpx_h_predictor_4x4_neon(uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -154,8 +150,7 @@ loop_v: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_h_predictor_4x4_neon: - vpx_h_predictor_4x4_neon: @ PROC +vpx_h_predictor_4x4_neon: @ PROC vld1.32 {d1[0]}, [r3] vdup.8 d0, d1[0] vst1.32 {d0[0]}, [r0], r1 @@ -166,7 +161,7 @@ _vpx_h_predictor_4x4_neon: vdup.8 d0, d1[3] vst1.32 {d0[0]}, [r0], r1 bx lr - .size vpx_h_predictor_4x4_neon, .-vpx_h_predictor_4x4_neon @ ENDP @ |vpx_h_predictor_4x4_neon| +.size vpx_h_predictor_4x4_neon, .-vpx_h_predictor_4x4_neon @ ENDP @ |vpx_h_predictor_4x4_neon| @void vpx_h_predictor_8x8_neon(uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -176,8 +171,7 @@ _vpx_h_predictor_4x4_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_h_predictor_8x8_neon: - vpx_h_predictor_8x8_neon: @ PROC +vpx_h_predictor_8x8_neon: @ PROC vld1.64 {d1}, [r3] vdup.8 d0, d1[0] vst1.64 {d0}, [r0], r1 @@ -196,7 +190,7 @@ _vpx_h_predictor_8x8_neon: vdup.8 d0, d1[7] vst1.64 {d0}, [r0], r1 bx lr - .size vpx_h_predictor_8x8_neon, .-vpx_h_predictor_8x8_neon @ ENDP @ |vpx_h_predictor_8x8_neon| +.size vpx_h_predictor_8x8_neon, .-vpx_h_predictor_8x8_neon @ ENDP @ |vpx_h_predictor_8x8_neon| @void vpx_h_predictor_16x16_neon(uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -206,8 +200,7 @@ _vpx_h_predictor_8x8_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_h_predictor_16x16_neon: - vpx_h_predictor_16x16_neon: @ PROC +vpx_h_predictor_16x16_neon: @ PROC vld1.8 {q1}, [r3] vdup.8 q0, d2[0] vst1.8 {q0}, [r0], r1 @@ -242,7 +235,7 @@ _vpx_h_predictor_16x16_neon: vdup.8 q0, d3[7] vst1.8 {q0}, [r0], r1 bx lr - .size vpx_h_predictor_16x16_neon, .-vpx_h_predictor_16x16_neon @ ENDP @ |vpx_h_predictor_16x16_neon| +.size vpx_h_predictor_16x16_neon, .-vpx_h_predictor_16x16_neon @ ENDP @ |vpx_h_predictor_16x16_neon| @void vpx_h_predictor_32x32_neon(uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -252,8 +245,7 @@ _vpx_h_predictor_16x16_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_h_predictor_32x32_neon: - vpx_h_predictor_32x32_neon: @ PROC +vpx_h_predictor_32x32_neon: @ PROC sub r1, r1, #16 mov r2, #2 loop_h: @@ -309,7 +301,7 @@ loop_h: subs r2, r2, #1 bgt loop_h bx lr - .size vpx_h_predictor_32x32_neon, .-vpx_h_predictor_32x32_neon @ ENDP @ |vpx_h_predictor_32x32_neon| +.size vpx_h_predictor_32x32_neon, .-vpx_h_predictor_32x32_neon @ ENDP @ |vpx_h_predictor_32x32_neon| @void vpx_tm_predictor_4x4_neon (uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -319,8 +311,7 @@ loop_h: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_tm_predictor_4x4_neon: - vpx_tm_predictor_4x4_neon: @ PROC +vpx_tm_predictor_4x4_neon: @ PROC @ Load ytop_left = above[-1]; sub r12, r2, #1 vld1.u8 {d0[]}, [r12] @@ -356,7 +347,7 @@ _vpx_tm_predictor_4x4_neon: vst1.32 {d0[0]}, [r0], r1 vst1.32 {d1[0]}, [r0], r1 bx lr - .size vpx_tm_predictor_4x4_neon, .-vpx_tm_predictor_4x4_neon @ ENDP @ |vpx_tm_predictor_4x4_neon| +.size vpx_tm_predictor_4x4_neon, .-vpx_tm_predictor_4x4_neon @ ENDP @ |vpx_tm_predictor_4x4_neon| @void vpx_tm_predictor_8x8_neon (uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -366,8 +357,7 @@ _vpx_tm_predictor_4x4_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_tm_predictor_8x8_neon: - vpx_tm_predictor_8x8_neon: @ PROC +vpx_tm_predictor_8x8_neon: @ PROC @ Load ytop_left = above[-1]; sub r12, r2, #1 vld1.8 {d0[]}, [r12] @@ -429,7 +419,7 @@ _vpx_tm_predictor_8x8_neon: vst1.64 {d3}, [r0], r1 bx lr - .size vpx_tm_predictor_8x8_neon, .-vpx_tm_predictor_8x8_neon @ ENDP @ |vpx_tm_predictor_8x8_neon| +.size vpx_tm_predictor_8x8_neon, .-vpx_tm_predictor_8x8_neon @ ENDP @ |vpx_tm_predictor_8x8_neon| @void vpx_tm_predictor_16x16_neon (uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -439,8 +429,7 @@ _vpx_tm_predictor_8x8_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_tm_predictor_16x16_neon: - vpx_tm_predictor_16x16_neon: @ PROC +vpx_tm_predictor_16x16_neon: @ PROC @ Load ytop_left = above[-1]; sub r12, r2, #1 vld1.8 {d0[]}, [r12] @@ -523,7 +512,7 @@ loop_16x16_neon: bgt loop_16x16_neon bx lr - .size vpx_tm_predictor_16x16_neon, .-vpx_tm_predictor_16x16_neon @ ENDP @ |vpx_tm_predictor_16x16_neon| +.size vpx_tm_predictor_16x16_neon, .-vpx_tm_predictor_16x16_neon @ ENDP @ |vpx_tm_predictor_16x16_neon| @void vpx_tm_predictor_32x32_neon (uint8_t *dst, ptrdiff_t y_stride, @ const uint8_t *above, @@ -533,8 +522,7 @@ loop_16x16_neon: @ r2 const uint8_t *above @ r3 const uint8_t *left -_vpx_tm_predictor_32x32_neon: - vpx_tm_predictor_32x32_neon: @ PROC +vpx_tm_predictor_32x32_neon: @ PROC @ Load ytop_left = above[-1]; sub r12, r2, #1 vld1.8 {d0[]}, [r12] @@ -653,6 +641,6 @@ loop_32x32_neon: bgt loop_32x32_neon bx lr - .size vpx_tm_predictor_32x32_neon, .-vpx_tm_predictor_32x32_neon @ ENDP @ |vpx_tm_predictor_32x32_neon| +.size vpx_tm_predictor_32x32_neon, .-vpx_tm_predictor_32x32_neon @ ENDP @ |vpx_tm_predictor_32x32_neon| - .section .note.GNU-stack,"",%progbits + .section .note.GNU-stack,"",%progbits |