diff options
Diffstat (limited to 'media/ffvpx/libavutil/aarch64')
| -rw-r--r-- | media/ffvpx/libavutil/aarch64/asm.S | 237 | ||||
| -rw-r--r-- | media/ffvpx/libavutil/aarch64/bswap.h | 56 | ||||
| -rw-r--r-- | media/ffvpx/libavutil/aarch64/cpu.c | 38 | ||||
| -rw-r--r-- | media/ffvpx/libavutil/aarch64/cpu.h | 29 | ||||
| -rw-r--r-- | media/ffvpx/libavutil/aarch64/float_dsp_init.c | 69 | ||||
| -rw-r--r-- | media/ffvpx/libavutil/aarch64/float_dsp_neon.S | 202 | ||||
| -rw-r--r-- | media/ffvpx/libavutil/aarch64/moz.build | 19 | ||||
| -rw-r--r-- | media/ffvpx/libavutil/aarch64/timer.h | 50 |
8 files changed, 700 insertions, 0 deletions
diff --git a/media/ffvpx/libavutil/aarch64/asm.S b/media/ffvpx/libavutil/aarch64/asm.S new file mode 100644 index 0000000000..a7782415d7 --- /dev/null +++ b/media/ffvpx/libavutil/aarch64/asm.S @@ -0,0 +1,237 @@ +/* + * Copyright (c) 2008 Mans Rullgard <mans@mansr.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 "config.h" + +#ifdef __ELF__ +# define ELF +#else +# define ELF # +#endif + +#if HAVE_AS_FUNC +# define FUNC +#else +# define FUNC # +#endif + +#ifndef __has_feature +# define __has_feature(x) 0 +#endif + + +/* Support macros for + * - Armv8.3-A Pointer Authentication and + * - Armv8.5-A Branch Target Identification + * features which require emitting a .note.gnu.property section with the + * appropriate architecture-dependent feature bits set. + * + * |AARCH64_SIGN_LINK_REGISTER| and |AARCH64_VALIDATE_LINK_REGISTER| expand to + * PACIxSP and AUTIxSP, respectively. |AARCH64_SIGN_LINK_REGISTER| should be + * used immediately before saving the LR register (x30) to the stack. + * |AARCH64_VALIDATE_LINK_REGISTER| should be used immediately after restoring + * it. Note |AARCH64_SIGN_LINK_REGISTER|'s modifications to LR must be undone + * with |AARCH64_VALIDATE_LINK_REGISTER| before RET. The SP register must also + * have the same value at the two points. For example: + * + * .global f + * f: + * AARCH64_SIGN_LINK_REGISTER + * stp x29, x30, [sp, #-96]! + * mov x29, sp + * ... + * ldp x29, x30, [sp], #96 + * AARCH64_VALIDATE_LINK_REGISTER + * ret + * + * |AARCH64_VALID_CALL_TARGET| expands to BTI 'c'. Either it, or + * |AARCH64_SIGN_LINK_REGISTER|, must be used at every point that may be an + * indirect call target. In particular, all symbols exported from a file must + * begin with one of these macros. For example, a leaf function that does not + * save LR can instead use |AARCH64_VALID_CALL_TARGET|: + * + * .globl return_zero + * return_zero: + * AARCH64_VALID_CALL_TARGET + * mov x0, #0 + * ret + * + * A non-leaf function which does not immediately save LR may need both macros + * because |AARCH64_SIGN_LINK_REGISTER| appears late. For example, the function + * may jump to an alternate implementation before setting up the stack: + * + * .globl with_early_jump + * with_early_jump: + * AARCH64_VALID_CALL_TARGET + * cmp x0, #128 + * b.lt .Lwith_early_jump_128 + * AARCH64_SIGN_LINK_REGISTER + * stp x29, x30, [sp, #-96]! + * mov x29, sp + * ... + * ldp x29, x30, [sp], #96 + * AARCH64_VALIDATE_LINK_REGISTER + * ret + * + * .Lwith_early_jump_128: + * ... + * ret + * + * These annotations are only required with indirect calls. Private symbols that + * are only the target of direct calls do not require annotations. Also note + * that |AARCH64_VALID_CALL_TARGET| is only valid for indirect calls (BLR), not + * indirect jumps (BR). Indirect jumps in assembly are supported through + * |AARCH64_VALID_JUMP_TARGET|. Landing Pads which shall serve for jumps and + * calls can be created using |AARCH64_VALID_JUMP_CALL_TARGET|. + * + * Although not necessary, it is safe to use these macros in 32-bit ARM + * assembly. This may be used to simplify dual 32-bit and 64-bit files. + * + * References: + * - "ELF for the Arm® 64-bit Architecture" + * https: *github.com/ARM-software/abi-aa/blob/master/aaelf64/aaelf64.rst + * - "Providing protection for complex software" + * https://developer.arm.com/architectures/learn-the-architecture/providing-protection-for-complex-software + */ +#if defined(__ARM_FEATURE_BTI_DEFAULT) && (__ARM_FEATURE_BTI_DEFAULT == 1) +# define GNU_PROPERTY_AARCH64_BTI (1 << 0) // Has BTI +# define AARCH64_VALID_CALL_TARGET hint #34 // BTI 'c' +# define AARCH64_VALID_JUMP_TARGET hint #38 // BTI 'j' +#else +# define GNU_PROPERTY_AARCH64_BTI 0 // No BTI +# define AARCH64_VALID_CALL_TARGET +# define AARCH64_VALID_JUMP_TARGET +#endif + +#if defined(__ARM_FEATURE_PAC_DEFAULT) +# if ((__ARM_FEATURE_PAC_DEFAULT & (1 << 0)) != 0) // authentication using key A +# define AARCH64_SIGN_LINK_REGISTER paciasp +# define AARCH64_VALIDATE_LINK_REGISTER autiasp +# elif ((__ARM_FEATURE_PAC_DEFAULT & (1 << 1)) != 0) // authentication using key B +# define AARCH64_SIGN_LINK_REGISTER pacibsp +# define AARCH64_VALIDATE_LINK_REGISTER autibsp +# else +# error Pointer authentication defines no valid key! +# endif +# if ((__ARM_FEATURE_PAC_DEFAULT & (1 << 2)) != 0) +# error Authentication of leaf functions is enabled but not supported in FFmpeg! +# endif +# define GNU_PROPERTY_AARCH64_PAC (1 << 1) +#else +# define GNU_PROPERTY_AARCH64_PAC 0 +# define AARCH64_SIGN_LINK_REGISTER +# define AARCH64_VALIDATE_LINK_REGISTER +#endif + + +#if (GNU_PROPERTY_AARCH64_BTI != 0 || GNU_PROPERTY_AARCH64_PAC != 0) && defined(__ELF__) + .pushsection .note.gnu.property, "a" + .balign 8 + .long 4 + .long 0x10 + .long 0x5 + .asciz "GNU" + .long 0xc0000000 /* GNU_PROPERTY_AARCH64_FEATURE_1_AND */ + .long 4 + .long (GNU_PROPERTY_AARCH64_BTI | GNU_PROPERTY_AARCH64_PAC) + .long 0 + .popsection +#endif + +.macro function name, export=0, align=2 + .macro endfunc +ELF .size \name, . - \name +FUNC .endfunc + .purgem endfunc + .endm + .text + .align \align + .if \export + .global EXTERN_ASM\name +ELF .type EXTERN_ASM\name, %function +FUNC .func EXTERN_ASM\name +EXTERN_ASM\name: + AARCH64_VALID_CALL_TARGET + .else +ELF .type \name, %function +FUNC .func \name +\name: + .endif +.endm + +.macro const name, align=2, relocate=0 + .macro endconst +ELF .size \name, . - \name + .purgem endconst + .endm +#if HAVE_SECTION_DATA_REL_RO +.if \relocate + .section .data.rel.ro +.else + .section .rodata +.endif +#elif defined(_WIN32) + .section .rdata +#elif !defined(__MACH__) + .section .rodata +#else + .const_data +#endif + .align \align +\name: +.endm + +.macro movrel rd, val, offset=0 +#if CONFIG_PIC && defined(__APPLE__) + .if \offset < 0 + adrp \rd, \val@PAGE + add \rd, \rd, \val@PAGEOFF + sub \rd, \rd, -(\offset) + .else + adrp \rd, \val+(\offset)@PAGE + add \rd, \rd, \val+(\offset)@PAGEOFF + .endif +#elif CONFIG_PIC && defined(_WIN32) + .if \offset < 0 + adrp \rd, \val + add \rd, \rd, :lo12:\val + sub \rd, \rd, -(\offset) + .else + adrp \rd, \val+(\offset) + add \rd, \rd, :lo12:\val+(\offset) + .endif +#elif CONFIG_PIC +# if __has_feature(hwaddress_sanitizer) + adrp \rd, :pg_hi21_nc:\val+(\offset) +# else + adrp \rd, \val+(\offset) +# endif + add \rd, \rd, :lo12:\val+(\offset) +#else + ldr \rd, =\val+\offset +#endif +.endm + +#define GLUE(a, b) a ## b +#define JOIN(a, b) GLUE(a, b) +#define X(s) JOIN(EXTERN_ASM, s) + +#define x18 do_not_use_x18 +#define w18 do_not_use_w18 diff --git a/media/ffvpx/libavutil/aarch64/bswap.h b/media/ffvpx/libavutil/aarch64/bswap.h new file mode 100644 index 0000000000..7abca657ba --- /dev/null +++ b/media/ffvpx/libavutil/aarch64/bswap.h @@ -0,0 +1,56 @@ +/* + * 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 AVUTIL_AARCH64_BSWAP_H +#define AVUTIL_AARCH64_BSWAP_H + +#include <stdint.h> +#include "config.h" +#include "libavutil/attributes.h" + +#if HAVE_INLINE_ASM + +#define av_bswap16 av_bswap16 +static av_always_inline av_const unsigned av_bswap16(unsigned x) +{ + unsigned y; + + __asm__("rev16 %w0, %w1" : "=r"(y) : "r"(x)); + return y; +} + +#define av_bswap32 av_bswap32 +static av_always_inline av_const uint32_t av_bswap32(uint32_t x) +{ + uint32_t y; + + __asm__("rev %w0, %w1" : "=r"(y) : "r"(x)); + return y; +} + +#define av_bswap64 av_bswap64 +static av_always_inline av_const uint64_t av_bswap64(uint64_t x) +{ + uint64_t y; + + __asm__("rev %0, %1" : "=r"(y) : "r"(x)); + return y; +} + +#endif /* HAVE_INLINE_ASM */ +#endif /* AVUTIL_AARCH64_BSWAP_H */ diff --git a/media/ffvpx/libavutil/aarch64/cpu.c b/media/ffvpx/libavutil/aarch64/cpu.c new file mode 100644 index 0000000000..cc641da576 --- /dev/null +++ b/media/ffvpx/libavutil/aarch64/cpu.c @@ -0,0 +1,38 @@ +/* + * 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 "libavutil/cpu.h" +#include "libavutil/cpu_internal.h" +#include "config.h" + +int ff_get_cpu_flags_aarch64(void) +{ + return AV_CPU_FLAG_ARMV8 * HAVE_ARMV8 | + AV_CPU_FLAG_NEON * HAVE_NEON | + AV_CPU_FLAG_VFP * HAVE_VFP; +} + +size_t ff_get_cpu_max_align_aarch64(void) +{ + int flags = av_get_cpu_flags(); + + if (flags & AV_CPU_FLAG_NEON) + return 16; + + return 8; +} diff --git a/media/ffvpx/libavutil/aarch64/cpu.h b/media/ffvpx/libavutil/aarch64/cpu.h new file mode 100644 index 0000000000..2ee3f9323a --- /dev/null +++ b/media/ffvpx/libavutil/aarch64/cpu.h @@ -0,0 +1,29 @@ +/* + * 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 AVUTIL_AARCH64_CPU_H +#define AVUTIL_AARCH64_CPU_H + +#include "libavutil/cpu.h" +#include "libavutil/cpu_internal.h" + +#define have_armv8(flags) CPUEXT(flags, ARMV8) +#define have_neon(flags) CPUEXT(flags, NEON) +#define have_vfp(flags) CPUEXT(flags, VFP) + +#endif /* AVUTIL_AARCH64_CPU_H */ diff --git a/media/ffvpx/libavutil/aarch64/float_dsp_init.c b/media/ffvpx/libavutil/aarch64/float_dsp_init.c new file mode 100644 index 0000000000..4325071821 --- /dev/null +++ b/media/ffvpx/libavutil/aarch64/float_dsp_init.c @@ -0,0 +1,69 @@ +/* + * ARM NEON optimised Float DSP functions + * Copyright (c) 2008 Mans Rullgard <mans@mansr.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 <stdint.h> + +#include "libavutil/attributes.h" +#include "libavutil/cpu.h" +#include "libavutil/float_dsp.h" +#include "cpu.h" + +void ff_vector_fmul_neon(float *dst, const float *src0, const float *src1, + int len); + +void ff_vector_fmac_scalar_neon(float *dst, const float *src, float mul, + int len); + +void ff_vector_fmul_scalar_neon(float *dst, const float *src, float mul, + int len); + +void ff_vector_dmul_scalar_neon(double *dst, const double *src, double mul, + int len); + +void ff_vector_fmul_window_neon(float *dst, const float *src0, + const float *src1, const float *win, int len); + +void ff_vector_fmul_add_neon(float *dst, const float *src0, const float *src1, + const float *src2, int len); + +void ff_vector_fmul_reverse_neon(float *dst, const float *src0, + const float *src1, int len); + +void ff_butterflies_float_neon(float *v1, float *v2, int len); + +float ff_scalarproduct_float_neon(const float *v1, const float *v2, int len); + +av_cold void ff_float_dsp_init_aarch64(AVFloatDSPContext *fdsp) +{ + int cpu_flags = av_get_cpu_flags(); + + if (have_neon(cpu_flags)) { + fdsp->butterflies_float = ff_butterflies_float_neon; + fdsp->scalarproduct_float = ff_scalarproduct_float_neon; + fdsp->vector_dmul_scalar = ff_vector_dmul_scalar_neon; + fdsp->vector_fmul = ff_vector_fmul_neon; + fdsp->vector_fmac_scalar = ff_vector_fmac_scalar_neon; + fdsp->vector_fmul_add = ff_vector_fmul_add_neon; + fdsp->vector_fmul_reverse = ff_vector_fmul_reverse_neon; + fdsp->vector_fmul_scalar = ff_vector_fmul_scalar_neon; + fdsp->vector_fmul_window = ff_vector_fmul_window_neon; + } +} diff --git a/media/ffvpx/libavutil/aarch64/float_dsp_neon.S b/media/ffvpx/libavutil/aarch64/float_dsp_neon.S new file mode 100644 index 0000000000..02d790c0cc --- /dev/null +++ b/media/ffvpx/libavutil/aarch64/float_dsp_neon.S @@ -0,0 +1,202 @@ +/* + * ARM NEON optimised Float DSP functions + * Copyright (c) 2008 Mans Rullgard <mans@mansr.com> + * Copyright (c) 2014 Janne Grunau <janne-libav@jannau.net> + * + * 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 "config.h" +#include "asm.S" + +function ff_vector_fmul_neon, export=1 +1: subs w3, w3, #16 + ld1 {v0.4S, v1.4S}, [x1], #32 + ld1 {v2.4S, v3.4S}, [x1], #32 + ld1 {v4.4S, v5.4S}, [x2], #32 + ld1 {v6.4S, v7.4S}, [x2], #32 + fmul v16.4S, v0.4S, v4.4S + fmul v17.4S, v1.4S, v5.4S + fmul v18.4S, v2.4S, v6.4S + fmul v19.4S, v3.4S, v7.4S + st1 {v16.4S, v17.4S}, [x0], #32 + st1 {v18.4S, v19.4S}, [x0], #32 + b.ne 1b + ret +endfunc + +function ff_vector_fmac_scalar_neon, export=1 + mov x3, #-32 +1: subs w2, w2, #16 + ld1 {v16.4S, v17.4S}, [x0], #32 + ld1 {v18.4S, v19.4S}, [x0], x3 + ld1 {v4.4S, v5.4S}, [x1], #32 + ld1 {v6.4S, v7.4S}, [x1], #32 + fmla v16.4S, v4.4S, v0.S[0] + fmla v17.4S, v5.4S, v0.S[0] + fmla v18.4S, v6.4S, v0.S[0] + fmla v19.4S, v7.4S, v0.S[0] + st1 {v16.4S, v17.4S}, [x0], #32 + st1 {v18.4S, v19.4S}, [x0], #32 + b.ne 1b + ret +endfunc + +function ff_vector_fmul_scalar_neon, export=1 + mov w4, #15 + bics w3, w2, w4 + dup v16.4S, v0.S[0] + b.eq 3f + ld1 {v0.4S, v1.4S}, [x1], #32 +1: subs w3, w3, #16 + fmul v0.4S, v0.4S, v16.4S + ld1 {v2.4S, v3.4S}, [x1], #32 + fmul v1.4S, v1.4S, v16.4S + fmul v2.4S, v2.4S, v16.4S + st1 {v0.4S, v1.4S}, [x0], #32 + fmul v3.4S, v3.4S, v16.4S + b.eq 2f + ld1 {v0.4S, v1.4S}, [x1], #32 + st1 {v2.4S, v3.4S}, [x0], #32 + b 1b +2: ands w2, w2, #15 + st1 {v2.4S, v3.4S}, [x0], #32 + b.eq 4f +3: ld1 {v0.4S}, [x1], #16 + fmul v0.4S, v0.4S, v16.4S + st1 {v0.4S}, [x0], #16 + subs w2, w2, #4 + b.gt 3b +4: ret +endfunc + +function ff_vector_dmul_scalar_neon, export=1 + dup v16.2D, v0.D[0] + ld1 {v0.2D, v1.2D}, [x1], #32 +1: subs w2, w2, #8 + fmul v0.2D, v0.2D, v16.2D + ld1 {v2.2D, v3.2D}, [x1], #32 + fmul v1.2D, v1.2D, v16.2D + fmul v2.2D, v2.2D, v16.2D + st1 {v0.2D, v1.2D}, [x0], #32 + fmul v3.2D, v3.2D, v16.2D + ld1 {v0.2D, v1.2D}, [x1], #32 + st1 {v2.2D, v3.2D}, [x0], #32 + b.gt 1b + ret +endfunc + +function ff_vector_fmul_window_neon, export=1 + sxtw x4, w4 // len + sub x2, x2, #8 + sub x5, x4, #2 + add x2, x2, x5, lsl #2 // src1 + 4 * (len - 4) + add x6, x3, x5, lsl #3 // win + 8 * (len - 2) + add x5, x0, x5, lsl #3 // dst + 8 * (len - 2) + mov x7, #-16 + ld1 {v0.4S}, [x1], #16 // s0 + ld1 {v2.4S}, [x3], #16 // wi + ld1 {v1.4S}, [x2], x7 // s1 +1: ld1 {v3.4S}, [x6], x7 // wj + subs x4, x4, #4 + fmul v17.4S, v0.4S, v2.4S // s0 * wi + rev64 v4.4S, v1.4S + rev64 v5.4S, v3.4S + rev64 v17.4S, v17.4S + ext v4.16B, v4.16B, v4.16B, #8 // s1_r + ext v5.16B, v5.16B, v5.16B, #8 // wj_r + ext v17.16B, v17.16B, v17.16B, #8 // (s0 * wi)_rev + fmul v16.4S, v0.4S, v5.4S // s0 * wj_r + fmla v17.4S, v1.4S, v3.4S // (s0 * wi)_rev + s1 * wj + b.eq 2f + ld1 {v0.4S}, [x1], #16 + fmls v16.4S, v4.4S, v2.4S // s0 * wj_r - s1_r * wi + st1 {v17.4S}, [x5], x7 + ld1 {v2.4S}, [x3], #16 + ld1 {v1.4S}, [x2], x7 + st1 {v16.4S}, [x0], #16 + b 1b +2: + fmls v16.4S, v4.4S, v2.4S // s0 * wj_r - s1_r * wi + st1 {v17.4S}, [x5], x7 + st1 {v16.4S}, [x0], #16 + ret +endfunc + +function ff_vector_fmul_add_neon, export=1 + ld1 {v0.4S, v1.4S}, [x1], #32 + ld1 {v2.4S, v3.4S}, [x2], #32 + ld1 {v4.4S, v5.4S}, [x3], #32 +1: subs w4, w4, #8 + fmla v4.4S, v0.4S, v2.4S + fmla v5.4S, v1.4S, v3.4S + b.eq 2f + ld1 {v0.4S, v1.4S}, [x1], #32 + ld1 {v2.4S, v3.4S}, [x2], #32 + st1 {v4.4S, v5.4S}, [x0], #32 + ld1 {v4.4S, v5.4S}, [x3], #32 + b 1b +2: st1 {v4.4S, v5.4S}, [x0], #32 + ret +endfunc + +function ff_vector_fmul_reverse_neon, export=1 + sxtw x3, w3 + add x2, x2, x3, lsl #2 + sub x2, x2, #32 + mov x4, #-32 + ld1 {v2.4S, v3.4S}, [x2], x4 + ld1 {v0.4S, v1.4S}, [x1], #32 +1: subs x3, x3, #8 + rev64 v3.4S, v3.4S + rev64 v2.4S, v2.4S + ext v3.16B, v3.16B, v3.16B, #8 + ext v2.16B, v2.16B, v2.16B, #8 + fmul v16.4S, v0.4S, v3.4S + fmul v17.4S, v1.4S, v2.4S + b.eq 2f + ld1 {v2.4S, v3.4S}, [x2], x4 + ld1 {v0.4S, v1.4S}, [x1], #32 + st1 {v16.4S, v17.4S}, [x0], #32 + b 1b +2: st1 {v16.4S, v17.4S}, [x0], #32 + ret +endfunc + +function ff_butterflies_float_neon, export=1 +1: ld1 {v0.4S}, [x0] + ld1 {v1.4S}, [x1] + subs w2, w2, #4 + fsub v2.4S, v0.4S, v1.4S + fadd v3.4S, v0.4S, v1.4S + st1 {v2.4S}, [x1], #16 + st1 {v3.4S}, [x0], #16 + b.gt 1b + ret +endfunc + +function ff_scalarproduct_float_neon, export=1 + movi v2.4S, #0 +1: ld1 {v0.4S}, [x0], #16 + ld1 {v1.4S}, [x1], #16 + subs w2, w2, #4 + fmla v2.4S, v0.4S, v1.4S + b.gt 1b + faddp v0.4S, v2.4S, v2.4S + faddp s0, v0.2S + ret +endfunc diff --git a/media/ffvpx/libavutil/aarch64/moz.build b/media/ffvpx/libavutil/aarch64/moz.build new file mode 100644 index 0000000000..9233abf768 --- /dev/null +++ b/media/ffvpx/libavutil/aarch64/moz.build @@ -0,0 +1,19 @@ +# -*- Mode: python; indent-tabs-mode: nil; tab-width: 40 -*- +# vim: set filetype=python: +# This Source Code Form is subject to the terms of the Mozilla Public +# License, v. 2.0. If a copy of the MPL was not distributed with this +# file, You can obtain one at http://mozilla.org/MPL/2.0/. + +SOURCES += [ + 'cpu.c', + 'float_dsp_init.c', + 'float_dsp_neon.S', +] + +if CONFIG['OS_ARCH'] == 'WINNT': + USE_INTEGRATED_CLANGCL_AS = True + DEFINES['EXTERN_ASM'] = '' + +FINAL_LIBRARY = 'mozavutil' + +include('/media/ffvpx/ffvpxcommon.mozbuild') diff --git a/media/ffvpx/libavutil/aarch64/timer.h b/media/ffvpx/libavutil/aarch64/timer.h new file mode 100644 index 0000000000..8b28fd354c --- /dev/null +++ b/media/ffvpx/libavutil/aarch64/timer.h @@ -0,0 +1,50 @@ +/* + * Copyright (c) 2015 Janne Grunau <janne-libav@jannau.net> + * + * 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 AVUTIL_AARCH64_TIMER_H +#define AVUTIL_AARCH64_TIMER_H + +#include <stdint.h> +#include "config.h" + +#if defined(__APPLE__) + +#include <mach/mach_time.h> + +#define AV_READ_TIME mach_absolute_time + +#elif HAVE_INLINE_ASM + +#define AV_READ_TIME read_time + +static inline uint64_t read_time(void) +{ + uint64_t cycle_counter; + __asm__ volatile( + "isb \t\n" + "mrs %0, pmccntr_el0 " + : "=r"(cycle_counter) :: "memory" ); + + return cycle_counter; +} + +#endif /* HAVE_INLINE_ASM */ + +#endif /* AVUTIL_AARCH64_TIMER_H */ |