1/**************************************************************************
2 *
3 * Copyright 2010 VMware, Inc.
4 * All Rights Reserved.
5 *
6 * Permission is hereby granted, free of charge, to any person obtaining a
7 * copy of this software and associated documentation files (the
8 * "Software"), to deal in the Software without restriction, including
9 * without limitation the rights to use, copy, modify, merge, publish,
10 * distribute, sub license, and/or sell copies of the Software, and to
11 * permit persons to whom the Software is furnished to do so, subject to
12 * the following conditions:
13 *
14 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
15 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
16 * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
17 * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
18 * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
19 * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
20 * USE OR OTHER DEALINGS IN THE SOFTWARE.
21 *
22 * The above copyright notice and this permission notice (including the
23 * next paragraph) shall be included in all copies or substantial portions
24 * of the Software.
25 *
26 **************************************************************************/
27
28
29#include "util/format/u_format_other.h"
30#include "util/u_math.h"
31#include "util/format_rgb9e5.h"
32#include "util/format_r11g11b10f.h"
33
34
35void
36util_format_r9g9b9e5_float_unpack_rgba_float(void *restrict dst_row,
37                                        const uint8_t *restrict src_row,
38                                        unsigned width)
39{
40   unsigned x;
41   float *dst = dst_row;
42   const uint8_t *src = src_row;
43   for(x = 0; x < width; x += 1) {
44      uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
45      rgb9e5_to_float3(value, dst);
46      dst[3] = 1; /* a */
47      src += 4;
48      dst += 4;
49   }
50}
51
52void
53util_format_r9g9b9e5_float_pack_rgba_float(uint8_t *restrict dst_row, unsigned dst_stride,
54                                      const float *restrict src_row, unsigned src_stride,
55                                      unsigned width, unsigned height)
56{
57   unsigned x, y;
58   for(y = 0; y < height; y += 1) {
59      const float *src = src_row;
60      uint8_t *dst = dst_row;
61      for(x = 0; x < width; x += 1) {
62         uint32_t value = util_cpu_to_le32(float3_to_rgb9e5(src));
63         *(uint32_t *)dst = value;
64         src += 4;
65         dst += 4;
66      }
67      dst_row += dst_stride;
68      src_row += src_stride/sizeof(*src_row);
69   }
70}
71
72void
73util_format_r9g9b9e5_float_fetch_rgba(void *restrict in_dst, const uint8_t *restrict src,
74                                       UNUSED unsigned i, UNUSED unsigned j)
75{
76   float *dst = in_dst;
77   uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
78   rgb9e5_to_float3(value, dst);
79   dst[3] = 1; /* a */
80}
81
82
83void
84util_format_r9g9b9e5_float_unpack_rgba_8unorm(uint8_t *restrict dst_row,
85                                         const uint8_t *restrict src_row,
86                                         unsigned width)
87{
88   unsigned x;
89   float p[3];
90   uint8_t *dst = dst_row;
91   const uint8_t *src = src_row;
92   for(x = 0; x < width; x += 1) {
93      uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
94      rgb9e5_to_float3(value, p);
95      dst[0] = float_to_ubyte(p[0]); /* r */
96      dst[1] = float_to_ubyte(p[1]); /* g */
97      dst[2] = float_to_ubyte(p[2]); /* b */
98      dst[3] = 255; /* a */
99      src += 4;
100      dst += 4;
101   }
102}
103
104
105void
106util_format_r9g9b9e5_float_pack_rgba_8unorm(uint8_t *restrict dst_row, unsigned dst_stride,
107                                       const uint8_t *restrict src_row, unsigned src_stride,
108                                       unsigned width, unsigned height)
109{
110   unsigned x, y;
111   float p[3];
112   for(y = 0; y < height; y += 1) {
113      const uint8_t *src = src_row;
114      uint8_t *dst = dst_row;
115      for(x = 0; x < width; x += 1) {
116         uint32_t value;
117         p[0] = ubyte_to_float(src[0]);
118         p[1] = ubyte_to_float(src[1]);
119         p[2] = ubyte_to_float(src[2]);
120         value = util_cpu_to_le32(float3_to_rgb9e5(p));
121         *(uint32_t *)dst = value;
122         src += 4;
123         dst += 4;
124      }
125      dst_row += dst_stride;
126      src_row += src_stride/sizeof(*src_row);
127   }
128}
129
130
131void
132util_format_r11g11b10_float_unpack_rgba_float(void *restrict dst_row,
133                                        const uint8_t *restrict src_row,
134                                        unsigned width)
135{
136   unsigned x;
137   float *dst = dst_row;
138   const uint8_t *src = src_row;
139   for(x = 0; x < width; x += 1) {
140      uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
141      r11g11b10f_to_float3(value, dst);
142      dst[3] = 1; /* a */
143      src += 4;
144      dst += 4;
145   }
146}
147
148void
149util_format_r11g11b10_float_pack_rgba_float(uint8_t *restrict dst_row, unsigned dst_stride,
150                                      const float *restrict src_row, unsigned src_stride,
151                                      unsigned width, unsigned height)
152{
153   unsigned x, y;
154   for(y = 0; y < height; y += 1) {
155      const float *src = src_row;
156      uint8_t *dst = dst_row;
157      for(x = 0; x < width; x += 1) {
158         uint32_t value = util_cpu_to_le32(float3_to_r11g11b10f(src));
159         *(uint32_t *)dst = value;
160         src += 4;
161         dst += 4;
162      }
163      dst_row += dst_stride;
164      src_row += src_stride/sizeof(*src_row);
165   }
166}
167
168void
169util_format_r11g11b10_float_fetch_rgba(void *restrict in_dst, const uint8_t *restrict src,
170                                        UNUSED unsigned i, UNUSED unsigned j)
171{
172   float *dst = in_dst;
173   uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
174   r11g11b10f_to_float3(value, dst);
175   dst[3] = 1; /* a */
176}
177
178
179void
180util_format_r11g11b10_float_unpack_rgba_8unorm(uint8_t *restrict dst_row,
181                                         const uint8_t *restrict src_row,
182                                         unsigned width)
183{
184   unsigned x;
185   float p[3];
186   uint8_t *dst = dst_row;
187   const uint8_t *src = src_row;
188   for(x = 0; x < width; x += 1) {
189      uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
190      r11g11b10f_to_float3(value, p);
191      dst[0] = float_to_ubyte(p[0]); /* r */
192      dst[1] = float_to_ubyte(p[1]); /* g */
193      dst[2] = float_to_ubyte(p[2]); /* b */
194      dst[3] = 255; /* a */
195      src += 4;
196      dst += 4;
197   }
198}
199
200
201void
202util_format_r11g11b10_float_pack_rgba_8unorm(uint8_t *restrict dst_row, unsigned dst_stride,
203                                       const uint8_t *restrict src_row, unsigned src_stride,
204                                       unsigned width, unsigned height)
205{
206   unsigned x, y;
207   float p[3];
208   for(y = 0; y < height; y += 1) {
209      const uint8_t *src = src_row;
210      uint8_t *dst = dst_row;
211      for(x = 0; x < width; x += 1) {
212         uint32_t value;
213         p[0] = ubyte_to_float(src[0]);
214         p[1] = ubyte_to_float(src[1]);
215         p[2] = ubyte_to_float(src[2]);
216         value = util_cpu_to_le32(float3_to_r11g11b10f(p));
217         *(uint32_t *)dst = value;
218         src += 4;
219         dst += 4;
220      }
221      dst_row += dst_stride;
222      src_row += src_stride/sizeof(*src_row);
223   }
224}
225
226/*
227 * PIPE_FORMAT_R8G8Bx_SNORM
228 *
229 * A.k.a. D3DFMT_CxV8U8
230 */
231
232static uint8_t
233r8g8bx_derive(int16_t r, int16_t g)
234{
235   /* Derive blue from red and green components.
236    * Apparently, we must always use integers to perform calculations,
237    * otherwise the results won't match D3D's CxV8U8 definition.
238    */
239   return (uint8_t)sqrtf(0x7f * 0x7f - r * r - g * g) * 0xff / 0x7f;
240}
241
242void
243util_format_r8g8bx_snorm_unpack_rgba_float(void *restrict dst_row,
244                                      const uint8_t *restrict src_row, unsigned width)
245{
246   unsigned x;
247   float *dst = dst_row;
248   const uint16_t *src = (const uint16_t *)src_row;
249   for(x = 0; x < width; x += 1) {
250      uint16_t value = util_cpu_to_le16(*src++);
251      int16_t r, g;
252
253      r = ((int16_t)(value << 8)) >> 8;
254      g = ((int16_t)(value << 0)) >> 8;
255
256      dst[0] = (float)(r * (1.0f/0x7f)); /* r */
257      dst[1] = (float)(g * (1.0f/0x7f)); /* g */
258      dst[2] = r8g8bx_derive(r, g) * (1.0f/0xff); /* b */
259      dst[3] = 1.0f; /* a */
260      dst += 4;
261   }
262}
263
264
265void
266util_format_r8g8bx_snorm_unpack_rgba_8unorm(uint8_t *restrict dst,
267                                       const uint8_t *restrict src_row,
268                                       unsigned width)
269{
270   unsigned x;
271   const uint16_t *src = (const uint16_t *)src_row;
272   for(x = 0; x < width; x += 1) {
273      uint16_t value = util_cpu_to_le16(*src++);
274      int16_t r, g;
275
276      r = ((int16_t)(value << 8)) >> 8;
277      g = ((int16_t)(value << 0)) >> 8;
278
279      dst[0] = (uint8_t)(((uint16_t)MAX2(r, 0)) * 0xff / 0x7f); /* r */
280      dst[1] = (uint8_t)(((uint16_t)MAX2(g, 0)) * 0xff / 0x7f); /* g */
281      dst[2] = r8g8bx_derive(r, g); /* b */
282      dst[3] = 255; /* a */
283      dst += 4;
284   }
285}
286
287
288void
289util_format_r8g8bx_snorm_pack_rgba_float(uint8_t *restrict dst_row, unsigned dst_stride,
290                                    const float *restrict src_row, unsigned src_stride,
291                                    unsigned width, unsigned height)
292{
293   unsigned x, y;
294   for(y = 0; y < height; y += 1) {
295      const float *src = src_row;
296      uint16_t *dst = (uint16_t *)dst_row;
297      for(x = 0; x < width; x += 1) {
298         uint16_t value = 0;
299
300         value |= (uint16_t)(((int8_t)(CLAMP(src[0], -1, 1) * 0x7f)) & 0xff) ;
301         value |= (uint16_t)((((int8_t)(CLAMP(src[1], -1, 1) * 0x7f)) & 0xff) << 8) ;
302
303         *dst++ = util_le16_to_cpu(value);
304
305         src += 4;
306      }
307      dst_row += dst_stride;
308      src_row += src_stride/sizeof(*src_row);
309   }
310}
311
312
313void
314util_format_r8g8bx_snorm_pack_rgba_8unorm(uint8_t *restrict dst_row, unsigned dst_stride,
315                                     const uint8_t *restrict src_row, unsigned src_stride,
316                                     unsigned width, unsigned height)
317{
318   unsigned x, y;
319
320   for(y = 0; y < height; y += 1) {
321      const uint8_t *src = src_row;
322      uint16_t *dst = (uint16_t *)dst_row;
323      for(x = 0; x < width; x += 1) {
324         uint16_t value = 0;
325
326         value |= src[0] >> 1;
327         value |= (src[1] >> 1) << 8;
328
329         *dst++ = util_le16_to_cpu(value);
330
331         src += 4;
332      }
333      dst_row += dst_stride;
334      src_row += src_stride/sizeof(*src_row);
335   }
336}
337
338
339void
340util_format_r8g8bx_snorm_fetch_rgba(void *restrict in_dst, const uint8_t *restrict src,
341                                     UNUSED unsigned i, UNUSED unsigned j)
342{
343   float *dst = in_dst;
344   uint16_t value = util_cpu_to_le16(*(const uint16_t *)src);
345   int16_t r, g;
346
347   r = ((int16_t)(value << 8)) >> 8;
348   g = ((int16_t)(value << 0)) >> 8;
349
350   dst[0] = r * (1.0f/0x7f); /* r */
351   dst[1] = g * (1.0f/0x7f); /* g */
352   dst[2] = r8g8bx_derive(r, g) * (1.0f/0xff); /* b */
353   dst[3] = 1.0f; /* a */
354}
355