| 1 |
/****************************************************************************** |
| 2 |
* Copyright © 2018, VideoLAN and dav1d authors |
| 3 |
* Copyright © 2015 Martin Storsjo |
| 4 |
* Copyright © 2015 Janne Grunau |
| 5 |
* All rights reserved. |
| 6 |
* |
| 7 |
* Redistribution and use in source and binary forms, with or without |
| 8 |
* modification, are permitted provided that the following conditions are met: |
| 9 |
* |
| 10 |
* 1. Redistributions of source code must retain the above copyright notice, this |
| 11 |
* list of conditions and the following disclaimer. |
| 12 |
* |
| 13 |
* 2. Redistributions in binary form must reproduce the above copyright notice, |
| 14 |
* this list of conditions and the following disclaimer in the documentation |
| 15 |
* and/or other materials provided with the distribution. |
| 16 |
* |
| 17 |
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND |
| 18 |
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED |
| 19 |
* WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
| 20 |
* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR |
| 21 |
* ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES |
| 22 |
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; |
| 23 |
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND |
| 24 |
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
| 25 |
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS |
| 26 |
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 27 |
*****************************************************************************/ |
| 28 |
|
| 29 |
#ifndef DAV1D_SRC_ARM_32_UTIL_S |
| 30 |
#define DAV1D_SRC_ARM_32_UTIL_S |
| 31 |
|
| 32 |
#include "config.h" |
| 33 |
#include "src/arm/asm.S" |
| 34 |
|
| 35 |
.macro movrel_local rd, val, offset=0 |
| 36 |
#if defined(PIC) |
| 37 |
ldr \rd, 90001f |
| 38 |
b 90002f |
| 39 |
90001: |
| 40 |
.word \val + \offset - (90002f + 8 - 4 * CONFIG_THUMB) |
| 41 |
90002: |
| 42 |
add \rd, \rd, pc |
| 43 |
#else |
| 44 |
movw \rd, #:lower16:\val+\offset |
| 45 |
movt \rd, #:upper16:\val+\offset |
| 46 |
#endif |
| 47 |
.endm |
| 48 |
|
| 49 |
.macro movrel rd, val, offset=0 |
| 50 |
#if defined(PIC) && defined(__APPLE__) |
| 51 |
ldr \rd, 1f |
| 52 |
b 2f |
| 53 |
1: |
| 54 |
.word 3f - (2f + 8 - 4 * CONFIG_THUMB) |
| 55 |
2: |
| 56 |
ldr \rd, [pc, \rd] |
| 57 |
.if \offset < 0 |
| 58 |
sub \rd, \rd, #-(\offset) |
| 59 |
.elseif \offset > 0 |
| 60 |
add \rd, \rd, #\offset |
| 61 |
.endif |
| 62 |
.non_lazy_symbol_pointer |
| 63 |
3: |
| 64 |
.indirect_symbol \val |
| 65 |
.word 0 |
| 66 |
.text |
| 67 |
#else |
| 68 |
movrel_local \rd, \val, \offset |
| 69 |
#endif |
| 70 |
.endm |
| 71 |
|
| 72 |
// This macro clobbers r7 (and r12 on windows) and stores data at the |
| 73 |
// bottom of the stack; sp is the start of the space allocated that |
| 74 |
// the caller can use. |
| 75 |
.macro sub_sp_align space |
| 76 |
#if CONFIG_THUMB |
| 77 |
mov r7, sp |
| 78 |
and r7, r7, #15 |
| 79 |
#else |
| 80 |
and r7, sp, #15 |
| 81 |
#endif |
| 82 |
sub sp, sp, r7 |
| 83 |
// Now the stack is aligned, store the amount of adjustment back |
| 84 |
// on the stack, as we don't want to waste a register as frame |
| 85 |
// pointer. |
| 86 |
str r7, [sp, #-16]! |
| 87 |
#ifdef _WIN32 |
| 88 |
.if \space > 8192 |
| 89 |
// Here, we'd need to touch two (or more) pages while decrementing |
| 90 |
// the stack pointer. |
| 91 |
.error "sub_sp_align doesn't support values over 8K at the moment" |
| 92 |
.elseif \space > 4096 |
| 93 |
sub r7, sp, #4096 |
| 94 |
ldr r12, [r7] |
| 95 |
sub r7, r7, #(\space - 4096) |
| 96 |
mov sp, r7 |
| 97 |
.else |
| 98 |
sub sp, sp, #\space |
| 99 |
.endif |
| 100 |
#else |
| 101 |
.if \space >= 4096 |
| 102 |
sub sp, sp, #(\space)/4096*4096 |
| 103 |
.endif |
| 104 |
.if (\space % 4096) != 0 |
| 105 |
sub sp, sp, #(\space)%4096 |
| 106 |
.endif |
| 107 |
#endif |
| 108 |
.endm |
| 109 |
|
| 110 |
.macro add_sp_align space |
| 111 |
.if \space >= 4096 |
| 112 |
add sp, sp, #(\space)/4096*4096 |
| 113 |
.endif |
| 114 |
.if (\space % 4096) != 0 |
| 115 |
add sp, sp, #(\space)%4096 |
| 116 |
.endif |
| 117 |
ldr r7, [sp], #16 |
| 118 |
// Add back the original stack adjustment |
| 119 |
add sp, sp, r7 |
| 120 |
.endm |
| 121 |
|
| 122 |
.macro transpose_8x8b q0, q1, q2, q3, r0, r1, r2, r3, r4, r5, r6, r7 |
| 123 |
vtrn.32 \q0, \q2 |
| 124 |
vtrn.32 \q1, \q3 |
| 125 |
|
| 126 |
vtrn.16 \r0, \r2 |
| 127 |
vtrn.16 \r1, \r3 |
| 128 |
vtrn.16 \r4, \r6 |
| 129 |
vtrn.16 \r5, \r7 |
| 130 |
|
| 131 |
vtrn.8 \r0, \r1 |
| 132 |
vtrn.8 \r2, \r3 |
| 133 |
vtrn.8 \r4, \r5 |
| 134 |
vtrn.8 \r6, \r7 |
| 135 |
.endm |
| 136 |
|
| 137 |
.macro transpose_8x8h r0, r1, r2, r3, r4, r5, r6, r7, d0, d1, d2, d3, d4, d5, d6, d7 |
| 138 |
vswp \d0, \d4 |
| 139 |
vswp \d1, \d5 |
| 140 |
vswp \d2, \d6 |
| 141 |
vswp \d3, \d7 |
| 142 |
|
| 143 |
vtrn.32 \r0, \r2 |
| 144 |
vtrn.32 \r1, \r3 |
| 145 |
vtrn.32 \r4, \r6 |
| 146 |
vtrn.32 \r5, \r7 |
| 147 |
|
| 148 |
vtrn.16 \r0, \r1 |
| 149 |
vtrn.16 \r2, \r3 |
| 150 |
vtrn.16 \r4, \r5 |
| 151 |
vtrn.16 \r6, \r7 |
| 152 |
.endm |
| 153 |
|
| 154 |
.macro transpose_4x8b q0, q1, r0, r1, r2, r3 |
| 155 |
vtrn.16 \q0, \q1 |
| 156 |
|
| 157 |
vtrn.8 \r0, \r1 |
| 158 |
vtrn.8 \r2, \r3 |
| 159 |
.endm |
| 160 |
|
| 161 |
.macro transpose_4x4s q0, q1, q2, q3, r0, r1, r2, r3, r4, r5, r6, r7 |
| 162 |
vswp \r1, \r4 // vtrn.64 \q0, \q2 |
| 163 |
vswp \r3, \r6 // vtrn.64 \q1, \q3 |
| 164 |
|
| 165 |
vtrn.32 \q0, \q1 |
| 166 |
vtrn.32 \q2, \q3 |
| 167 |
.endm |
| 168 |
|
| 169 |
.macro transpose_4x4h q0, q1, r0, r1, r2, r3 |
| 170 |
vtrn.32 \q0, \q1 |
| 171 |
|
| 172 |
vtrn.16 \r0, \r1 |
| 173 |
vtrn.16 \r2, \r3 |
| 174 |
.endm |
| 175 |
|
| 176 |
.macro transpose_4x8h r0, r1, r2, r3 |
| 177 |
vtrn.32 \r0, \r2 |
| 178 |
vtrn.32 \r1, \r3 |
| 179 |
|
| 180 |
vtrn.16 \r0, \r1 |
| 181 |
vtrn.16 \r2, \r3 |
| 182 |
.endm |
| 183 |
|
| 184 |
#endif /* DAV1D_SRC_ARM_32_UTIL_S */ |