| 1 | /* |
| 2 | * Fast C2P (Chunky-to-Planar) Conversion |
| 3 | * |
| 4 | * NOTES: |
| 5 | * - This code was inspired by Scout's C2P tutorial |
| 6 | * - It assumes to run on a big endian system |
| 7 | * |
| 8 | * Copyright © 2003-2008 Geert Uytterhoeven |
| 9 | * |
| 10 | * Permission is hereby granted, free of charge, to any person obtaining a |
| 11 | * copy of this software and associated documentation files (the "Software"), |
| 12 | * to deal in the Software without restriction, including without limitation |
| 13 | * the rights to use, copy, modify, merge, publish, distribute, sublicense, |
| 14 | * and/or sell copies of the Software, and to permit persons to whom the |
| 15 | * Software is furnished to do so, subject to the following conditions: |
| 16 | * |
| 17 | * The above copyright notice and this permission notice (including the next |
| 18 | * paragraph) shall be included in all copies or substantial portions of the |
| 19 | * Software. |
| 20 | * |
| 21 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| 22 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 23 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL |
| 24 | * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
| 25 | * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING |
| 26 | * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER |
| 27 | * DEALINGS IN THE SOFTWARE. |
| 28 | */ |
| 29 | |
| 30 | |
| 31 | /* |
| 32 | * Basic transpose step |
| 33 | */ |
| 34 | |
| 35 | static inline void _transp(CARD32 d[], unsigned int i1, unsigned int i2, |
| 36 | unsigned int shift, CARD32 mask) |
| 37 | { |
| 38 | CARD32 t = (d[i1] ^ (d[i2] >> shift)) & mask; |
| 39 | |
| 40 | d[i1] ^= t; |
| 41 | d[i2] ^= t << shift; |
| 42 | } |
| 43 | |
| 44 | |
| 45 | static inline void c2p_unsupported(void) { |
| 46 | BUG_WARN(1); |
| 47 | } |
| 48 | |
| 49 | static inline CARD32 get_mask(unsigned int n) |
| 50 | { |
| 51 | switch (n) { |
| 52 | case 1: |
| 53 | return 0x55555555; |
| 54 | |
| 55 | case 2: |
| 56 | return 0x33333333; |
| 57 | |
| 58 | case 4: |
| 59 | return 0x0f0f0f0f; |
| 60 | |
| 61 | case 8: |
| 62 | return 0x00ff00ff; |
| 63 | |
| 64 | case 16: |
| 65 | return 0x0000ffff; |
| 66 | } |
| 67 | |
| 68 | c2p_unsupported(); |
| 69 | return 0; |
| 70 | } |
| 71 | |
| 72 | |
| 73 | /* |
| 74 | * Transpose operations on 8 32-bit words |
| 75 | */ |
| 76 | |
| 77 | static inline void transp8(CARD32 d[], unsigned int n, unsigned int m) |
| 78 | { |
| 79 | CARD32 mask = get_mask(n); |
| 80 | |
| 81 | switch (m) { |
| 82 | case 1: |
| 83 | /* First n x 1 block */ |
| 84 | _transp(d, 0, 1, n, mask); |
| 85 | /* Second n x 1 block */ |
| 86 | _transp(d, 2, 3, n, mask); |
| 87 | /* Third n x 1 block */ |
| 88 | _transp(d, 4, 5, n, mask); |
| 89 | /* Fourth n x 1 block */ |
| 90 | _transp(d, 6, 7, n, mask); |
| 91 | return; |
| 92 | |
| 93 | case 2: |
| 94 | /* First n x 2 block */ |
| 95 | _transp(d, 0, 2, n, mask); |
| 96 | _transp(d, 1, 3, n, mask); |
| 97 | /* Second n x 2 block */ |
| 98 | _transp(d, 4, 6, n, mask); |
| 99 | _transp(d, 5, 7, n, mask); |
| 100 | return; |
| 101 | |
| 102 | case 4: |
| 103 | /* Single n x 4 block */ |
| 104 | _transp(d, 0, 4, n, mask); |
| 105 | _transp(d, 1, 5, n, mask); |
| 106 | _transp(d, 2, 6, n, mask); |
| 107 | _transp(d, 3, 7, n, mask); |
| 108 | return; |
| 109 | } |
| 110 | |
| 111 | c2p_unsupported(); |
| 112 | } |
| 113 | |
| 114 | |
| 115 | /* |
| 116 | * Transpose operations on 4 32-bit words |
| 117 | */ |
| 118 | |
| 119 | static inline void transp4(CARD32 d[], unsigned int n, unsigned int m) |
| 120 | { |
| 121 | CARD32 mask = get_mask(n); |
| 122 | |
| 123 | switch (m) { |
| 124 | case 1: |
| 125 | /* First n x 1 block */ |
| 126 | _transp(d, 0, 1, n, mask); |
| 127 | /* Second n x 1 block */ |
| 128 | _transp(d, 2, 3, n, mask); |
| 129 | return; |
| 130 | |
| 131 | case 2: |
| 132 | /* Single n x 2 block */ |
| 133 | _transp(d, 0, 2, n, mask); |
| 134 | _transp(d, 1, 3, n, mask); |
| 135 | return; |
| 136 | } |
| 137 | |
| 138 | c2p_unsupported(); |
| 139 | } |
| 140 | |
| 141 | |
| 142 | /* |
| 143 | * Transpose operations on 4 32-bit words (reverse order) |
| 144 | */ |
| 145 | |
| 146 | static inline void transp4x(CARD32 d[], unsigned int n, unsigned int m) |
| 147 | { |
| 148 | CARD32 mask = get_mask(n); |
| 149 | |
| 150 | switch (m) { |
| 151 | case 2: |
| 152 | /* Single n x 2 block */ |
| 153 | _transp(d, 2, 0, n, mask); |
| 154 | _transp(d, 3, 1, n, mask); |
| 155 | return; |
| 156 | } |
| 157 | |
| 158 | c2p_unsupported(); |
| 159 | } |
| 160 | |
| 161 | |
| 162 | /* |
| 163 | * Transpose operations on 2 32-bit words |
| 164 | */ |
| 165 | |
| 166 | static inline void transp2(CARD32 d[], unsigned int n) |
| 167 | { |
| 168 | CARD32 mask = get_mask(n); |
| 169 | |
| 170 | /* Single n x 1 block */ |
| 171 | _transp(d, 0, 1, n, mask); |
| 172 | return; |
| 173 | } |
| 174 | |
| 175 | |
| 176 | /* |
| 177 | * Transpose operations on 2 32-bit words (reverse order) |
| 178 | */ |
| 179 | |
| 180 | static inline void transp2x(CARD32 d[], unsigned int n) |
| 181 | { |
| 182 | CARD32 mask = get_mask(n); |
| 183 | |
| 184 | /* Single n x 1 block */ |
| 185 | _transp(d, 1, 0, n, mask); |
| 186 | return; |
| 187 | } |