diff options
Diffstat (limited to 'libavcodec/arm')
| -rw-r--r-- | libavcodec/arm/dcadsp_init_arm.c | 2 | ||||
| -rw-r--r-- | libavcodec/arm/dcadsp_neon.S | 5 | ||||
| -rw-r--r-- | libavcodec/arm/dsputil_init_neon.c | 3 | ||||
| -rw-r--r-- | libavcodec/arm/dsputil_neon.S | 15 | ||||
| -rw-r--r-- | libavcodec/arm/fft_init_arm.c | 2 | ||||
| -rw-r--r-- | libavcodec/arm/synth_filter_neon.S | 8 |
6 files changed, 13 insertions, 22 deletions
diff --git a/libavcodec/arm/dcadsp_init_arm.c b/libavcodec/arm/dcadsp_init_arm.c index 816718d483..5663cd7fc2 100644 --- a/libavcodec/arm/dcadsp_init_arm.c +++ b/libavcodec/arm/dcadsp_init_arm.c @@ -23,7 +23,7 @@ #include "libavcodec/dcadsp.h" void ff_dca_lfe_fir_neon(float *out, const float *in, const float *coefs, - int decifactor, float scale, float bias); + int decifactor, float scale); void av_cold ff_dcadsp_init_arm(DCADSPContext *s) { diff --git a/libavcodec/arm/dcadsp_neon.S b/libavcodec/arm/dcadsp_neon.S index 19960ab193..c3bddd3e41 100644 --- a/libavcodec/arm/dcadsp_neon.S +++ b/libavcodec/arm/dcadsp_neon.S @@ -29,7 +29,7 @@ function ff_dca_lfe_fir_neon, export=1 cmp r3, #32 moveq r6, #256/32 movne r6, #256/64 -NOVFP vldr d0, [sp, #16] @ scale, bias +NOVFP vldr s0, [sp, #16] @ scale mov lr, #-16 1: vmov.f32 q2, #0.0 @ v0 @@ -51,8 +51,7 @@ NOVFP vldr d0, [sp, #16] @ scale, bias vadd.f32 d4, d4, d5 vadd.f32 d6, d6, d7 vpadd.f32 d4, d4, d6 - vdup.32 d5, d0[1] - vmla.f32 d5, d4, d0[0] + vmul.f32 d5, d4, d0[0] vst1.32 {d5[0]}, [r0,:32]! vst1.32 {d5[1]}, [r4,:32]! bne 1b diff --git a/libavcodec/arm/dsputil_init_neon.c b/libavcodec/arm/dsputil_init_neon.c index 221183cef8..67982048f9 100644 --- a/libavcodec/arm/dsputil_init_neon.c +++ b/libavcodec/arm/dsputil_init_neon.c @@ -140,8 +140,7 @@ void ff_vp3_h_loop_filter_neon(uint8_t *, int, int *); void ff_vector_fmul_neon(float *dst, const float *src0, const float *src1, int len); void ff_vector_fmul_window_neon(float *dst, const float *src0, - const float *src1, const float *win, - float add_bias, int len); + const float *src1, const float *win, int len); void ff_vector_fmul_scalar_neon(float *dst, const float *src, float mul, int len); void ff_vector_fmul_sv_scalar_2_neon(float *dst, const float *src, diff --git a/libavcodec/arm/dsputil_neon.S b/libavcodec/arm/dsputil_neon.S index 42fb38de52..8329f6cc57 100644 --- a/libavcodec/arm/dsputil_neon.S +++ b/libavcodec/arm/dsputil_neon.S @@ -777,11 +777,8 @@ function ff_vector_fmul_neon, export=1 endfunc function ff_vector_fmul_window_neon, export=1 -VFP vdup.32 q8, d0[0] -NOVFP vld1.32 {d16[],d17[]}, [sp,:32] push {r4,r5,lr} -VFP ldr lr, [sp, #12] -NOVFP ldr lr, [sp, #16] + ldr lr, [sp, #12] sub r2, r2, #8 sub r5, lr, #2 add r2, r2, r5, lsl #2 @@ -793,14 +790,12 @@ NOVFP ldr lr, [sp, #16] vld1.64 {d4,d5}, [r3,:128]! vld1.64 {d6,d7}, [r4,:128], r5 1: subs lr, lr, #4 - vmov q11, q8 - vmla.f32 d22, d0, d4 - vmov q10, q8 - vmla.f32 d23, d1, d5 + vmul.f32 d22, d0, d4 vrev64.32 q3, q3 - vmla.f32 d20, d0, d7 + vmul.f32 d23, d1, d5 vrev64.32 q1, q1 - vmla.f32 d21, d1, d6 + vmul.f32 d20, d0, d7 + vmul.f32 d21, d1, d6 beq 2f vmla.f32 d22, d3, d7 vld1.64 {d0,d1}, [r1,:128]! diff --git a/libavcodec/arm/fft_init_arm.c b/libavcodec/arm/fft_init_arm.c index bde12400de..3f2d554413 100644 --- a/libavcodec/arm/fft_init_arm.c +++ b/libavcodec/arm/fft_init_arm.c @@ -34,7 +34,7 @@ void ff_synth_filter_float_neon(FFTContext *imdct, float *synth_buf_ptr, int *synth_buf_offset, float synth_buf2[32], const float window[512], float out[32], const float in[32], - float scale, float bias); + float scale); av_cold void ff_fft_init_arm(FFTContext *s) { diff --git a/libavcodec/arm/synth_filter_neon.S b/libavcodec/arm/synth_filter_neon.S index a7c23df02b..1464abe562 100644 --- a/libavcodec/arm/synth_filter_neon.S +++ b/libavcodec/arm/synth_filter_neon.S @@ -42,7 +42,7 @@ VFP vpop {d0} ldr r5, [sp, #9*4] @ window ldr r2, [sp, #10*4] @ out -NOVFP vldr d0, [sp, #12*4] @ scale, bias +NOVFP vldr s0, [sp, #12*4] @ scale add r8, r9, #12*4 mov lr, #64*4 @@ -90,10 +90,8 @@ NOVFP vldr d0, [sp, #12*4] @ scale, bias sub r11, r11, #512*4 b 2b 3: - vdup.32 q8, d0[1] - vdup.32 q9, d0[1] - vmla.f32 q8, q10, d0[0] - vmla.f32 q9, q1, d0[0] + vmul.f32 q8, q10, d0[0] + vmul.f32 q9, q1, d0[0] vst1.32 {q3}, [r3,:128] sub r3, r3, #16*4 vst1.32 {q2}, [r3,:128] |