diff options
author | marha <marha@users.sourceforge.net> | 2011-01-19 17:29:52 +0000 |
---|---|---|
committer | marha <marha@users.sourceforge.net> | 2011-01-19 17:29:52 +0000 |
commit | a13b75f056f9f9efcf6ecb8610b40ddbbb2bbb69 (patch) | |
tree | 778ce0682518f7a0615ce5585410f3aaecb14421 /pixman/pixman | |
parent | 6e3cfc5bc8ca969856e4d56dec01870df709d75a (diff) | |
download | vcxsrv-a13b75f056f9f9efcf6ecb8610b40ddbbb2bbb69.tar.gz vcxsrv-a13b75f056f9f9efcf6ecb8610b40ddbbb2bbb69.tar.bz2 vcxsrv-a13b75f056f9f9efcf6ecb8610b40ddbbb2bbb69.zip |
xserver pixman mesa git update 19 jan 2011
Diffstat (limited to 'pixman/pixman')
-rw-r--r-- | pixman/pixman/pixman-arm-common.h | 635 | ||||
-rw-r--r-- | pixman/pixman/pixman-bits-image.c | 235 | ||||
-rw-r--r-- | pixman/pixman/pixman-conical-gradient.c | 411 | ||||
-rw-r--r-- | pixman/pixman/pixman-fast-path.c | 22 | ||||
-rw-r--r-- | pixman/pixman/pixman-general.c | 628 | ||||
-rw-r--r-- | pixman/pixman/pixman-image.c | 1562 | ||||
-rw-r--r-- | pixman/pixman/pixman-implementation.c | 81 | ||||
-rw-r--r-- | pixman/pixman/pixman-linear-gradient.c | 552 | ||||
-rw-r--r-- | pixman/pixman/pixman-mmx.c | 6756 | ||||
-rw-r--r-- | pixman/pixman/pixman-private.h | 1779 | ||||
-rw-r--r-- | pixman/pixman/pixman-radial-gradient.c | 910 | ||||
-rw-r--r-- | pixman/pixman/pixman-solid-fill.c | 209 | ||||
-rw-r--r-- | pixman/pixman/pixman-sse2.c | 30 | ||||
-rw-r--r-- | pixman/pixman/pixman-utils.c | 6 |
14 files changed, 7043 insertions, 6773 deletions
diff --git a/pixman/pixman/pixman-arm-common.h b/pixman/pixman/pixman-arm-common.h index ede63a629..372e9f9a8 100644 --- a/pixman/pixman/pixman-arm-common.h +++ b/pixman/pixman/pixman-arm-common.h @@ -1,316 +1,319 @@ -/*
- * Copyright © 2010 Nokia Corporation
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
- * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
- * DEALINGS IN THE SOFTWARE.
- *
- * Author: Siarhei Siamashka (siarhei.siamashka@nokia.com)
- */
-
-#ifndef PIXMAN_ARM_COMMON_H
-#define PIXMAN_ARM_COMMON_H
-
-#include "pixman-fast-path.h"
-
-/* Define some macros which can expand into proxy functions between
- * ARM assembly optimized functions and the rest of pixman fast path API.
- *
- * All the low level ARM assembly functions have to use ARM EABI
- * calling convention and take up to 8 arguments:
- * width, height, dst, dst_stride, src, src_stride, mask, mask_stride
- *
- * The arguments are ordered with the most important coming first (the
- * first 4 arguments are passed to function in registers, the rest are
- * on stack). The last arguments are optional, for example if the
- * function is not using mask, then 'mask' and 'mask_stride' can be
- * omitted when doing a function call.
- *
- * Arguments 'src' and 'mask' contain either a pointer to the top left
- * pixel of the composited rectangle or a pixel color value depending
- * on the function type. In the case of just a color value (solid source
- * or mask), the corresponding stride argument is unused.
- */
-
-#define SKIP_ZERO_SRC 1
-#define SKIP_ZERO_MASK 2
-
-#define PIXMAN_ARM_BIND_FAST_PATH_SRC_DST(cputype, name, \
- src_type, src_cnt, \
- dst_type, dst_cnt) \
-void \
-pixman_composite_##name##_asm_##cputype (int32_t w, \
- int32_t h, \
- dst_type *dst, \
- int32_t dst_stride, \
- src_type *src, \
- int32_t src_stride); \
- \
-static void \
-cputype##_composite_##name (pixman_implementation_t *imp, \
- pixman_op_t op, \
- pixman_image_t * src_image, \
- pixman_image_t * mask_image, \
- pixman_image_t * dst_image, \
- int32_t src_x, \
- int32_t src_y, \
- int32_t mask_x, \
- int32_t mask_y, \
- int32_t dest_x, \
- int32_t dest_y, \
- int32_t width, \
- int32_t height) \
-{ \
- dst_type *dst_line; \
- src_type *src_line; \
- int32_t dst_stride, src_stride; \
- \
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, src_type, \
- src_stride, src_line, src_cnt); \
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \
- dst_stride, dst_line, dst_cnt); \
- \
- pixman_composite_##name##_asm_##cputype (width, height, \
- dst_line, dst_stride, \
- src_line, src_stride); \
-}
-
-#define PIXMAN_ARM_BIND_FAST_PATH_N_DST(flags, cputype, name, \
- dst_type, dst_cnt) \
-void \
-pixman_composite_##name##_asm_##cputype (int32_t w, \
- int32_t h, \
- dst_type *dst, \
- int32_t dst_stride, \
- uint32_t src); \
- \
-static void \
-cputype##_composite_##name (pixman_implementation_t *imp, \
- pixman_op_t op, \
- pixman_image_t * src_image, \
- pixman_image_t * mask_image, \
- pixman_image_t * dst_image, \
- int32_t src_x, \
- int32_t src_y, \
- int32_t mask_x, \
- int32_t mask_y, \
- int32_t dest_x, \
- int32_t dest_y, \
- int32_t width, \
- int32_t height) \
-{ \
- dst_type *dst_line; \
- int32_t dst_stride; \
- uint32_t src; \
- \
- src = _pixman_image_get_solid (src_image, dst_image->bits.format); \
- \
- if ((flags & SKIP_ZERO_SRC) && src == 0) \
- return; \
- \
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \
- dst_stride, dst_line, dst_cnt); \
- \
- pixman_composite_##name##_asm_##cputype (width, height, \
- dst_line, dst_stride, \
- src); \
-}
-
-#define PIXMAN_ARM_BIND_FAST_PATH_N_MASK_DST(flags, cputype, name, \
- mask_type, mask_cnt, \
- dst_type, dst_cnt) \
-void \
-pixman_composite_##name##_asm_##cputype (int32_t w, \
- int32_t h, \
- dst_type *dst, \
- int32_t dst_stride, \
- uint32_t src, \
- int32_t unused, \
- mask_type *mask, \
- int32_t mask_stride); \
- \
-static void \
-cputype##_composite_##name (pixman_implementation_t *imp, \
- pixman_op_t op, \
- pixman_image_t * src_image, \
- pixman_image_t * mask_image, \
- pixman_image_t * dst_image, \
- int32_t src_x, \
- int32_t src_y, \
- int32_t mask_x, \
- int32_t mask_y, \
- int32_t dest_x, \
- int32_t dest_y, \
- int32_t width, \
- int32_t height) \
-{ \
- dst_type *dst_line; \
- mask_type *mask_line; \
- int32_t dst_stride, mask_stride; \
- uint32_t src; \
- \
- src = _pixman_image_get_solid (src_image, dst_image->bits.format); \
- \
- if ((flags & SKIP_ZERO_SRC) && src == 0) \
- return; \
- \
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \
- dst_stride, dst_line, dst_cnt); \
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, mask_type, \
- mask_stride, mask_line, mask_cnt); \
- \
- pixman_composite_##name##_asm_##cputype (width, height, \
- dst_line, dst_stride, \
- src, 0, \
- mask_line, mask_stride); \
-}
-
-#define PIXMAN_ARM_BIND_FAST_PATH_SRC_N_DST(flags, cputype, name, \
- src_type, src_cnt, \
- dst_type, dst_cnt) \
-void \
-pixman_composite_##name##_asm_##cputype (int32_t w, \
- int32_t h, \
- dst_type *dst, \
- int32_t dst_stride, \
- src_type *src, \
- int32_t src_stride, \
- uint32_t mask); \
- \
-static void \
-cputype##_composite_##name (pixman_implementation_t *imp, \
- pixman_op_t op, \
- pixman_image_t * src_image, \
- pixman_image_t * mask_image, \
- pixman_image_t * dst_image, \
- int32_t src_x, \
- int32_t src_y, \
- int32_t mask_x, \
- int32_t mask_y, \
- int32_t dest_x, \
- int32_t dest_y, \
- int32_t width, \
- int32_t height) \
-{ \
- dst_type *dst_line; \
- src_type *src_line; \
- int32_t dst_stride, src_stride; \
- uint32_t mask; \
- \
- mask = _pixman_image_get_solid (mask_image, dst_image->bits.format);\
- \
- if ((flags & SKIP_ZERO_MASK) && mask == 0) \
- return; \
- \
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \
- dst_stride, dst_line, dst_cnt); \
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, src_type, \
- src_stride, src_line, src_cnt); \
- \
- pixman_composite_##name##_asm_##cputype (width, height, \
- dst_line, dst_stride, \
- src_line, src_stride, \
- mask); \
-}
-
-#define PIXMAN_ARM_BIND_FAST_PATH_SRC_MASK_DST(cputype, name, \
- src_type, src_cnt, \
- mask_type, mask_cnt, \
- dst_type, dst_cnt) \
-void \
-pixman_composite_##name##_asm_##cputype (int32_t w, \
- int32_t h, \
- dst_type *dst, \
- int32_t dst_stride, \
- src_type *src, \
- int32_t src_stride, \
- mask_type *mask, \
- int32_t mask_stride); \
- \
-static void \
-cputype##_composite_##name (pixman_implementation_t *imp, \
- pixman_op_t op, \
- pixman_image_t * src_image, \
- pixman_image_t * mask_image, \
- pixman_image_t * dst_image, \
- int32_t src_x, \
- int32_t src_y, \
- int32_t mask_x, \
- int32_t mask_y, \
- int32_t dest_x, \
- int32_t dest_y, \
- int32_t width, \
- int32_t height) \
-{ \
- dst_type *dst_line; \
- src_type *src_line; \
- mask_type *mask_line; \
- int32_t dst_stride, src_stride, mask_stride; \
- \
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \
- dst_stride, dst_line, dst_cnt); \
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, src_type, \
- src_stride, src_line, src_cnt); \
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, mask_type, \
- mask_stride, mask_line, mask_cnt); \
- \
- pixman_composite_##name##_asm_##cputype (width, height, \
- dst_line, dst_stride, \
- src_line, src_stride, \
- mask_line, mask_stride); \
-}
-
-#define PIXMAN_ARM_BIND_SCALED_NEAREST_SRC_DST(cputype, name, op, \
- src_type, dst_type) \
-void \
-pixman_scaled_nearest_scanline_##name##_##op##_asm_##cputype ( \
- int32_t w, \
- dst_type * dst, \
- src_type * src, \
- pixman_fixed_t vx, \
- pixman_fixed_t unit_x);\
- \
-static force_inline void \
-scaled_nearest_scanline_##cputype##_##name##_##op (dst_type * pd, \
- src_type * ps, \
- int32_t w, \
- pixman_fixed_t vx, \
- pixman_fixed_t unit_x, \
- pixman_fixed_t max_vx) \
-{ \
- pixman_scaled_nearest_scanline_##name##_##op##_asm_##cputype (w, pd, ps, \
- vx, unit_x);\
-} \
- \
-FAST_NEAREST_MAINLOOP (cputype##_##name##_cover_##op, \
- scaled_nearest_scanline_##cputype##_##name##_##op, \
- src_type, dst_type, COVER) \
-FAST_NEAREST_MAINLOOP (cputype##_##name##_none_##op, \
- scaled_nearest_scanline_##cputype##_##name##_##op, \
- src_type, dst_type, NONE) \
-FAST_NEAREST_MAINLOOP (cputype##_##name##_pad_##op, \
- scaled_nearest_scanline_##cputype##_##name##_##op, \
- src_type, dst_type, PAD)
-
-/* Provide entries for the fast path table */
-#define PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH(op,s,d,func) \
- SIMPLE_NEAREST_FAST_PATH_COVER (op,s,d,func), \
- SIMPLE_NEAREST_FAST_PATH_NONE (op,s,d,func), \
- SIMPLE_NEAREST_FAST_PATH_PAD (op,s,d,func)
-
-#endif
+/* + * Copyright © 2010 Nokia Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Author: Siarhei Siamashka (siarhei.siamashka@nokia.com) + */ + +#ifndef PIXMAN_ARM_COMMON_H +#define PIXMAN_ARM_COMMON_H + +#include "pixman-fast-path.h" + +/* Define some macros which can expand into proxy functions between + * ARM assembly optimized functions and the rest of pixman fast path API. + * + * All the low level ARM assembly functions have to use ARM EABI + * calling convention and take up to 8 arguments: + * width, height, dst, dst_stride, src, src_stride, mask, mask_stride + * + * The arguments are ordered with the most important coming first (the + * first 4 arguments are passed to function in registers, the rest are + * on stack). The last arguments are optional, for example if the + * function is not using mask, then 'mask' and 'mask_stride' can be + * omitted when doing a function call. + * + * Arguments 'src' and 'mask' contain either a pointer to the top left + * pixel of the composited rectangle or a pixel color value depending + * on the function type. In the case of just a color value (solid source + * or mask), the corresponding stride argument is unused. + */ + +#define SKIP_ZERO_SRC 1 +#define SKIP_ZERO_MASK 2 + +#define PIXMAN_ARM_BIND_FAST_PATH_SRC_DST(cputype, name, \ + src_type, src_cnt, \ + dst_type, dst_cnt) \ +void \ +pixman_composite_##name##_asm_##cputype (int32_t w, \ + int32_t h, \ + dst_type *dst, \ + int32_t dst_stride, \ + src_type *src, \ + int32_t src_stride); \ + \ +static void \ +cputype##_composite_##name (pixman_implementation_t *imp, \ + pixman_op_t op, \ + pixman_image_t * src_image, \ + pixman_image_t * mask_image, \ + pixman_image_t * dst_image, \ + int32_t src_x, \ + int32_t src_y, \ + int32_t mask_x, \ + int32_t mask_y, \ + int32_t dest_x, \ + int32_t dest_y, \ + int32_t width, \ + int32_t height) \ +{ \ + dst_type *dst_line; \ + src_type *src_line; \ + int32_t dst_stride, src_stride; \ + \ + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, src_type, \ + src_stride, src_line, src_cnt); \ + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \ + dst_stride, dst_line, dst_cnt); \ + \ + pixman_composite_##name##_asm_##cputype (width, height, \ + dst_line, dst_stride, \ + src_line, src_stride); \ +} + +#define PIXMAN_ARM_BIND_FAST_PATH_N_DST(flags, cputype, name, \ + dst_type, dst_cnt) \ +void \ +pixman_composite_##name##_asm_##cputype (int32_t w, \ + int32_t h, \ + dst_type *dst, \ + int32_t dst_stride, \ + uint32_t src); \ + \ +static void \ +cputype##_composite_##name (pixman_implementation_t *imp, \ + pixman_op_t op, \ + pixman_image_t * src_image, \ + pixman_image_t * mask_image, \ + pixman_image_t * dst_image, \ + int32_t src_x, \ + int32_t src_y, \ + int32_t mask_x, \ + int32_t mask_y, \ + int32_t dest_x, \ + int32_t dest_y, \ + int32_t width, \ + int32_t height) \ +{ \ + dst_type *dst_line; \ + int32_t dst_stride; \ + uint32_t src; \ + \ + src = _pixman_image_get_solid ( \ + imp, src_image, dst_image->bits.format); \ + \ + if ((flags & SKIP_ZERO_SRC) && src == 0) \ + return; \ + \ + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \ + dst_stride, dst_line, dst_cnt); \ + \ + pixman_composite_##name##_asm_##cputype (width, height, \ + dst_line, dst_stride, \ + src); \ +} + +#define PIXMAN_ARM_BIND_FAST_PATH_N_MASK_DST(flags, cputype, name, \ + mask_type, mask_cnt, \ + dst_type, dst_cnt) \ +void \ +pixman_composite_##name##_asm_##cputype (int32_t w, \ + int32_t h, \ + dst_type *dst, \ + int32_t dst_stride, \ + uint32_t src, \ + int32_t unused, \ + mask_type *mask, \ + int32_t mask_stride); \ + \ +static void \ +cputype##_composite_##name (pixman_implementation_t *imp, \ + pixman_op_t op, \ + pixman_image_t * src_image, \ + pixman_image_t * mask_image, \ + pixman_image_t * dst_image, \ + int32_t src_x, \ + int32_t src_y, \ + int32_t mask_x, \ + int32_t mask_y, \ + int32_t dest_x, \ + int32_t dest_y, \ + int32_t width, \ + int32_t height) \ +{ \ + dst_type *dst_line; \ + mask_type *mask_line; \ + int32_t dst_stride, mask_stride; \ + uint32_t src; \ + \ + src = _pixman_image_get_solid ( \ + imp, src_image, dst_image->bits.format); \ + \ + if ((flags & SKIP_ZERO_SRC) && src == 0) \ + return; \ + \ + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \ + dst_stride, dst_line, dst_cnt); \ + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, mask_type, \ + mask_stride, mask_line, mask_cnt); \ + \ + pixman_composite_##name##_asm_##cputype (width, height, \ + dst_line, dst_stride, \ + src, 0, \ + mask_line, mask_stride); \ +} + +#define PIXMAN_ARM_BIND_FAST_PATH_SRC_N_DST(flags, cputype, name, \ + src_type, src_cnt, \ + dst_type, dst_cnt) \ +void \ +pixman_composite_##name##_asm_##cputype (int32_t w, \ + int32_t h, \ + dst_type *dst, \ + int32_t dst_stride, \ + src_type *src, \ + int32_t src_stride, \ + uint32_t mask); \ + \ +static void \ +cputype##_composite_##name (pixman_implementation_t *imp, \ + pixman_op_t op, \ + pixman_image_t * src_image, \ + pixman_image_t * mask_image, \ + pixman_image_t * dst_image, \ + int32_t src_x, \ + int32_t src_y, \ + int32_t mask_x, \ + int32_t mask_y, \ + int32_t dest_x, \ + int32_t dest_y, \ + int32_t width, \ + int32_t height) \ +{ \ + dst_type *dst_line; \ + src_type *src_line; \ + int32_t dst_stride, src_stride; \ + uint32_t mask; \ + \ + mask = _pixman_image_get_solid ( \ + imp, mask_image, dst_image->bits.format); \ + \ + if ((flags & SKIP_ZERO_MASK) && mask == 0) \ + return; \ + \ + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \ + dst_stride, dst_line, dst_cnt); \ + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, src_type, \ + src_stride, src_line, src_cnt); \ + \ + pixman_composite_##name##_asm_##cputype (width, height, \ + dst_line, dst_stride, \ + src_line, src_stride, \ + mask); \ +} + +#define PIXMAN_ARM_BIND_FAST_PATH_SRC_MASK_DST(cputype, name, \ + src_type, src_cnt, \ + mask_type, mask_cnt, \ + dst_type, dst_cnt) \ +void \ +pixman_composite_##name##_asm_##cputype (int32_t w, \ + int32_t h, \ + dst_type *dst, \ + int32_t dst_stride, \ + src_type *src, \ + int32_t src_stride, \ + mask_type *mask, \ + int32_t mask_stride); \ + \ +static void \ +cputype##_composite_##name (pixman_implementation_t *imp, \ + pixman_op_t op, \ + pixman_image_t * src_image, \ + pixman_image_t * mask_image, \ + pixman_image_t * dst_image, \ + int32_t src_x, \ + int32_t src_y, \ + int32_t mask_x, \ + int32_t mask_y, \ + int32_t dest_x, \ + int32_t dest_y, \ + int32_t width, \ + int32_t height) \ +{ \ + dst_type *dst_line; \ + src_type *src_line; \ + mask_type *mask_line; \ + int32_t dst_stride, src_stride, mask_stride; \ + \ + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, dst_type, \ + dst_stride, dst_line, dst_cnt); \ + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, src_type, \ + src_stride, src_line, src_cnt); \ + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, mask_type, \ + mask_stride, mask_line, mask_cnt); \ + \ + pixman_composite_##name##_asm_##cputype (width, height, \ + dst_line, dst_stride, \ + src_line, src_stride, \ + mask_line, mask_stride); \ +} + +#define PIXMAN_ARM_BIND_SCALED_NEAREST_SRC_DST(cputype, name, op, \ + src_type, dst_type) \ +void \ +pixman_scaled_nearest_scanline_##name##_##op##_asm_##cputype ( \ + int32_t w, \ + dst_type * dst, \ + src_type * src, \ + pixman_fixed_t vx, \ + pixman_fixed_t unit_x);\ + \ +static force_inline void \ +scaled_nearest_scanline_##cputype##_##name##_##op (dst_type * pd, \ + src_type * ps, \ + int32_t w, \ + pixman_fixed_t vx, \ + pixman_fixed_t unit_x, \ + pixman_fixed_t max_vx) \ +{ \ + pixman_scaled_nearest_scanline_##name##_##op##_asm_##cputype (w, pd, ps, \ + vx, unit_x);\ +} \ + \ +FAST_NEAREST_MAINLOOP (cputype##_##name##_cover_##op, \ + scaled_nearest_scanline_##cputype##_##name##_##op, \ + src_type, dst_type, COVER) \ +FAST_NEAREST_MAINLOOP (cputype##_##name##_none_##op, \ + scaled_nearest_scanline_##cputype##_##name##_##op, \ + src_type, dst_type, NONE) \ +FAST_NEAREST_MAINLOOP (cputype##_##name##_pad_##op, \ + scaled_nearest_scanline_##cputype##_##name##_##op, \ + src_type, dst_type, PAD) + +/* Provide entries for the fast path table */ +#define PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH(op,s,d,func) \ + SIMPLE_NEAREST_FAST_PATH_COVER (op,s,d,func), \ + SIMPLE_NEAREST_FAST_PATH_NONE (op,s,d,func), \ + SIMPLE_NEAREST_FAST_PATH_PAD (op,s,d,func) + +#endif diff --git a/pixman/pixman/pixman-bits-image.c b/pixman/pixman/pixman-bits-image.c index c453e0ee6..a865d719a 100644 --- a/pixman/pixman/pixman-bits-image.c +++ b/pixman/pixman/pixman-bits-image.c @@ -35,43 +35,41 @@ #include "pixman-private.h" #include "pixman-combine32.h" -/* Store functions */ -void -_pixman_image_store_scanline_32 (bits_image_t * image, - int x, - int y, - int width, - const uint32_t *buffer) +/* + * By default, just evaluate the image at 32bpp and expand. Individual image + * types can plug in a better scanline getter if they want to. For example + * we could produce smoother gradients by evaluating them at higher color + * depth, but that's a project for the future. + */ +static void +_pixman_image_get_scanline_generic_64 (pixman_image_t * image, + int x, + int y, + int width, + uint32_t * buffer, + const uint32_t * mask) { - image->store_scanline_32 (image, x, y, width, buffer); + uint32_t *mask8 = NULL; - if (image->common.alpha_map) + /* Contract the mask image, if one exists, so that the 32-bit fetch + * function can use it. + */ + if (mask) { - x -= image->common.alpha_origin_x; - y -= image->common.alpha_origin_y; + mask8 = pixman_malloc_ab (width, sizeof(uint32_t)); + if (!mask8) + return; - image->common.alpha_map->store_scanline_32 ( - image->common.alpha_map, x, y, width, buffer); + pixman_contract (mask8, (uint64_t *)mask, width); } -} -void -_pixman_image_store_scanline_64 (bits_image_t * image, - int x, - int y, - int width, - const uint32_t *buffer) -{ - image->store_scanline_64 (image, x, y, width, buffer); + /* Fetch the source image into the first half of buffer. */ + image->bits.get_scanline_32 (image, x, y, width, (uint32_t*)buffer, mask8); - if (image->common.alpha_map) - { - x -= image->common.alpha_origin_x; - y -= image->common.alpha_origin_y; + /* Expand from 32bpp to 64bpp in place. */ + pixman_expand ((uint64_t *)buffer, buffer, PIXMAN_a8r8g8b8, width); - image->common.alpha_map->store_scanline_64 ( - image->common.alpha_map, x, y, width, buffer); - } + free (mask8); } /* Fetch functions */ @@ -297,6 +295,7 @@ bits_image_fetch_bilinear_no_repeat_8888 (pixman_image_t * ima, uint32_t *bottom_row; uint32_t *end; uint32_t zero[2] = { 0, 0 }; + uint32_t one = 1; int y, y1, y2; int disty; int mask_inc; @@ -362,10 +361,8 @@ bits_image_fetch_bilinear_no_repeat_8888 (pixman_image_t * ima, */ if (!mask) { - uint32_t mask_bits = 1; - mask_inc = 0; - mask = &mask_bits; + mask = &one; } else { @@ -1337,8 +1334,8 @@ bits_image_property_changed (pixman_image_t *image) if ((info->format == format || info->format == PIXMAN_any) && (info->flags & flags) == info->flags) { - image->common.get_scanline_32 = info->fetch_32; - image->common.get_scanline_64 = info->fetch_64; + image->bits.get_scanline_32 = info->fetch_32; + image->bits.get_scanline_64 = info->fetch_64; break; } @@ -1347,6 +1344,176 @@ bits_image_property_changed (pixman_image_t *image) } static uint32_t * +src_get_scanline_narrow (pixman_iter_t *iter, const uint32_t *mask) +{ + iter->image->bits.get_scanline_32 ( + iter->image, iter->x, iter->y++, iter->width, iter->buffer, mask); + + return iter->buffer; +} + +static uint32_t * +src_get_scanline_wide (pixman_iter_t *iter, const uint32_t *mask) +{ + iter->image->bits.get_scanline_64 ( + iter->image, iter->x, iter->y++, iter->width, iter->buffer, mask); + + return iter->buffer; +} + +void +_pixman_bits_image_src_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags) +{ + if (flags & ITER_NARROW) + iter->get_scanline = src_get_scanline_narrow; + else + iter->get_scanline = src_get_scanline_wide; +} + +static uint32_t * +dest_get_scanline_narrow (pixman_iter_t *iter, const uint32_t *mask) +{ + pixman_image_t *image = iter->image; + int x = iter->x; + int y = iter->y; + int width = iter->width; + uint32_t * buffer = iter->buffer; + + image->bits.fetch_scanline_32 (image, x, y, width, buffer, mask); + if (image->common.alpha_map) + { + x -= image->common.alpha_origin_x; + y -= image->common.alpha_origin_y; + + image->common.alpha_map->fetch_scanline_32 ( + (pixman_image_t *)image->common.alpha_map, + x, y, width, buffer, mask); + } + + return iter->buffer; +} + +static uint32_t * +dest_get_scanline_wide (pixman_iter_t *iter, const uint32_t *mask) +{ + bits_image_t * image = &iter->image->bits; + int x = iter->x; + int y = iter->y; + int width = iter->width; + uint32_t * buffer = iter->buffer; + + image->fetch_scanline_64 ( + (pixman_image_t *)image, x, y, width, buffer, mask); + if (image->common.alpha_map) + { + x -= image->common.alpha_origin_x; + y -= image->common.alpha_origin_y; + + image->common.alpha_map->fetch_scanline_64 ( + (pixman_image_t *)image->common.alpha_map, x, y, width, buffer, mask); + } + + return iter->buffer; +} + +static void +dest_write_back_narrow (pixman_iter_t *iter) +{ + bits_image_t * image = &iter->image->bits; + int x = iter->x; + int y = iter->y; + int width = iter->width; + const uint32_t *buffer = iter->buffer; + + image->store_scanline_32 (image, x, y, width, buffer); + + if (image->common.alpha_map) + { + x -= image->common.alpha_origin_x; + y -= image->common.alpha_origin_y; + + image->common.alpha_map->store_scanline_32 ( + image->common.alpha_map, x, y, width, buffer); + } + + iter->y++; +} + +static void +dest_write_back_wide (pixman_iter_t *iter) +{ + bits_image_t * image = &iter->image->bits; + int x = iter->x; + int y = iter->y; + int width = iter->width; + const uint32_t *buffer = iter->buffer; + + image->store_scanline_64 (image, x, y, width, buffer); + + if (image->common.alpha_map) + { + x -= image->common.alpha_origin_x; + y -= image->common.alpha_origin_y; + + image->common.alpha_map->store_scanline_64 ( + image->common.alpha_map, x, y, width, buffer); + } + + iter->y++; +} + +static void +dest_write_back_direct (pixman_iter_t *iter) +{ + iter->buffer += iter->image->bits.rowstride; +} + +void +_pixman_bits_image_dest_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags) +{ + if (flags & ITER_NARROW) + { + if (((image->common.flags & + (FAST_PATH_NO_ALPHA_MAP | FAST_PATH_NO_ACCESSORS)) == + (FAST_PATH_NO_ALPHA_MAP | FAST_PATH_NO_ACCESSORS)) && + (image->bits.format == PIXMAN_a8r8g8b8 || + (image->bits.format == PIXMAN_x8r8g8b8 && + (flags & ITER_LOCALIZED_ALPHA)))) + { + iter->buffer = image->bits.bits + y * image->bits.rowstride + x; + + iter->get_scanline = _pixman_iter_get_scanline_noop; + iter->write_back = dest_write_back_direct; + } + else + { + if ((flags & (ITER_IGNORE_RGB | ITER_IGNORE_ALPHA)) == + (ITER_IGNORE_RGB | ITER_IGNORE_ALPHA)) + { + iter->get_scanline = _pixman_iter_get_scanline_noop; + } + else + { + iter->get_scanline = dest_get_scanline_narrow; + } + + iter->write_back = dest_write_back_narrow; + } + } + else + { + iter->get_scanline = dest_get_scanline_wide; + iter->write_back = dest_write_back_wide; + } +} + +static uint32_t * create_bits (pixman_format_code_t format, int width, int height, diff --git a/pixman/pixman/pixman-conical-gradient.c b/pixman/pixman/pixman-conical-gradient.c index 769b984d1..9d7d2e8b5 100644 --- a/pixman/pixman/pixman-conical-gradient.c +++ b/pixman/pixman/pixman-conical-gradient.c @@ -1,197 +1,214 @@ -/*
- * Copyright © 2000 SuSE, Inc.
- * Copyright © 2007 Red Hat, Inc.
- * Copyright © 2000 Keith Packard, member of The XFree86 Project, Inc.
- * 2005 Lars Knoll & Zack Rusin, Trolltech
- *
- * Permission to use, copy, modify, distribute, and sell this software and its
- * documentation for any purpose is hereby granted without fee, provided that
- * the above copyright notice appear in all copies and that both that
- * copyright notice and this permission notice appear in supporting
- * documentation, and that the name of Keith Packard not be used in
- * advertising or publicity pertaining to distribution of the software without
- * specific, written prior permission. Keith Packard makes no
- * representations about the suitability of this software for any purpose. It
- * is provided "as is" without express or implied warranty.
- *
- * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
- * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
- * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
- * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING
- * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
- * SOFTWARE.
- */
-
-#ifdef HAVE_CONFIG_H
-#include <config.h>
-#endif
-
-#include <stdlib.h>
-#include <math.h>
-#include "pixman-private.h"
-
-static force_inline double
-coordinates_to_parameter (double x, double y, double angle)
-{
- double t;
-
- t = atan2 (y, x) + angle;
-
- while (t < 0)
- t += 2 * M_PI;
-
- while (t >= 2 * M_PI)
- t -= 2 * M_PI;
-
- return 1 - t * (1 / (2 * M_PI)); /* Scale t to [0, 1] and
- * make rotation CCW
- */
-}
-
-static void
-conical_gradient_get_scanline_32 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *mask)
-{
- gradient_t *gradient = (gradient_t *)image;
- conical_gradient_t *conical = (conical_gradient_t *)image;
- uint32_t *end = buffer + width;
- pixman_gradient_walker_t walker;
- pixman_bool_t affine = TRUE;
- double cx = 1.;
- double cy = 0.;
- double cz = 0.;
- double rx = x + 0.5;
- double ry = y + 0.5;
- double rz = 1.;
-
- _pixman_gradient_walker_init (&walker, gradient, image->common.repeat);
-
- if (image->common.transform)
- {
- pixman_vector_t v;
-
- /* reference point is the center of the pixel */
- v.vector[0] = pixman_int_to_fixed (x) + pixman_fixed_1 / 2;
- v.vector[1] = pixman_int_to_fixed (y) + pixman_fixed_1 / 2;
- v.vector[2] = pixman_fixed_1;
-
- if (!pixman_transform_point_3d (image->common.transform, &v))
- return;
-
- cx = image->common.transform->matrix[0][0] / 65536.;
- cy = image->common.transform->matrix[1][0] / 65536.;
- cz = image->common.transform->matrix[2][0] / 65536.;
-
- rx = v.vector[0] / 65536.;
- ry = v.vector[1] / 65536.;
- rz = v.vector[2] / 65536.;
-
- affine =
- image->common.transform->matrix[2][0] == 0 &&
- v.vector[2] == pixman_fixed_1;
- }
-
- if (affine)
- {
- rx -= conical->center.x / 65536.;
- ry -= conical->center.y / 65536.;
-
- while (buffer < end)
- {
- if (!mask || *mask++)
- {
- double t = coordinates_to_parameter (rx, ry, conical->angle);
-
- *buffer = _pixman_gradient_walker_pixel (
- &walker, (pixman_fixed_48_16_t)pixman_double_to_fixed (t));
- }
-
- ++buffer;
-
- rx += cx;
- ry += cy;
- }
- }
- else
- {
- while (buffer < end)
- {
- double x, y;
-
- if (!mask || *mask++)
- {
- double t;
-
- if (rz != 0)
- {
- x = rx / rz;
- y = ry / rz;
- }
- else
- {
- x = y = 0.;
- }
-
- x -= conical->center.x / 65536.;
- y -= conical->center.y / 65536.;
-
- t = coordinates_to_parameter (x, y, conical->angle);
-
- *buffer = _pixman_gradient_walker_pixel (
- &walker, (pixman_fixed_48_16_t)pixman_double_to_fixed (t));
- }
-
- ++buffer;
-
- rx += cx;
- ry += cy;
- rz += cz;
- }
- }
-}
-
-static void
-conical_gradient_property_changed (pixman_image_t *image)
-{
- image->common.get_scanline_32 = conical_gradient_get_scanline_32;
- image->common.get_scanline_64 = _pixman_image_get_scanline_generic_64;
-}
-
-PIXMAN_EXPORT pixman_image_t *
-pixman_image_create_conical_gradient (pixman_point_fixed_t * center,
- pixman_fixed_t angle,
- const pixman_gradient_stop_t *stops,
- int n_stops)
-{
- pixman_image_t *image = _pixman_image_allocate ();
- conical_gradient_t *conical;
-
- if (!image)
- return NULL;
-
- conical = &image->conical;
-
- if (!_pixman_init_gradient (&conical->common, stops, n_stops))
- {
- free (image);
- return NULL;
- }
-
- angle = MOD (angle, pixman_int_to_fixed (360));
-
- image->type = CONICAL;
-
- conical->center = *center;
- conical->angle = (pixman_fixed_to_double (angle) / 180.0) * M_PI;
-
- image->common.property_changed = conical_gradient_property_changed;
-
- return image;
-}
-
+/* + * Copyright © 2000 SuSE, Inc. + * Copyright © 2007 Red Hat, Inc. + * Copyright © 2000 Keith Packard, member of The XFree86 Project, Inc. + * 2005 Lars Knoll & Zack Rusin, Trolltech + * + * Permission to use, copy, modify, distribute, and sell this software and its + * documentation for any purpose is hereby granted without fee, provided that + * the above copyright notice appear in all copies and that both that + * copyright notice and this permission notice appear in supporting + * documentation, and that the name of Keith Packard not be used in + * advertising or publicity pertaining to distribution of the software without + * specific, written prior permission. Keith Packard makes no + * representations about the suitability of this software for any purpose. It + * is provided "as is" without express or implied warranty. + * + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS + * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND + * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY + * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN + * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING + * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS + * SOFTWARE. + */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#include <stdlib.h> +#include <math.h> +#include "pixman-private.h" + +static force_inline double +coordinates_to_parameter (double x, double y, double angle) +{ + double t; + + t = atan2 (y, x) + angle; + + while (t < 0) + t += 2 * M_PI; + + while (t >= 2 * M_PI) + t -= 2 * M_PI; + + return 1 - t * (1 / (2 * M_PI)); /* Scale t to [0, 1] and + * make rotation CCW + */ +} + +static uint32_t * +conical_get_scanline_narrow (pixman_iter_t *iter, const uint32_t *mask) +{ + pixman_image_t *image = iter->image; + int x = iter->x; + int y = iter->y; + int width = iter->width; + uint32_t *buffer = iter->buffer; + + gradient_t *gradient = (gradient_t *)image; + conical_gradient_t *conical = (conical_gradient_t *)image; + uint32_t *end = buffer + width; + pixman_gradient_walker_t walker; + pixman_bool_t affine = TRUE; + double cx = 1.; + double cy = 0.; + double cz = 0.; + double rx = x + 0.5; + double ry = y + 0.5; + double rz = 1.; + + _pixman_gradient_walker_init (&walker, gradient, image->common.repeat); + + if (image->common.transform) + { + pixman_vector_t v; + + /* reference point is the center of the pixel */ + v.vector[0] = pixman_int_to_fixed (x) + pixman_fixed_1 / 2; + v.vector[1] = pixman_int_to_fixed (y) + pixman_fixed_1 / 2; + v.vector[2] = pixman_fixed_1; + + if (!pixman_transform_point_3d (image->common.transform, &v)) + return iter->buffer; + + cx = image->common.transform->matrix[0][0] / 65536.; + cy = image->common.transform->matrix[1][0] / 65536.; + cz = image->common.transform->matrix[2][0] / 65536.; + + rx = v.vector[0] / 65536.; + ry = v.vector[1] / 65536.; + rz = v.vector[2] / 65536.; + + affine = + image->common.transform->matrix[2][0] == 0 && + v.vector[2] == pixman_fixed_1; + } + + if (affine) + { + rx -= conical->center.x / 65536.; + ry -= conical->center.y / 65536.; + + while (buffer < end) + { + if (!mask || *mask++) + { + double t = coordinates_to_parameter (rx, ry, conical->angle); + + *buffer = _pixman_gradient_walker_pixel ( + &walker, (pixman_fixed_48_16_t)pixman_double_to_fixed (t)); + } + + ++buffer; + + rx += cx; + ry += cy; + } + } + else + { + while (buffer < end) + { + double x, y; + + if (!mask || *mask++) + { + double t; + + if (rz != 0) + { + x = rx / rz; + y = ry / rz; + } + else + { + x = y = 0.; + } + + x -= conical->center.x / 65536.; + y -= conical->center.y / 65536.; + + t = coordinates_to_parameter (x, y, conical->angle); + + *buffer = _pixman_gradient_walker_pixel ( + &walker, (pixman_fixed_48_16_t)pixman_double_to_fixed (t)); + } + + ++buffer; + + rx += cx; + ry += cy; + rz += cz; + } + } + + iter->y++; + return iter->buffer; +} + +static uint32_t * +conical_get_scanline_wide (pixman_iter_t *iter, const uint32_t *mask) +{ + uint32_t *buffer = conical_get_scanline_narrow (iter, NULL); + + pixman_expand ((uint64_t *)buffer, buffer, PIXMAN_a8r8g8b8, iter->width); + + return buffer; +} + +void +_pixman_conical_gradient_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags) +{ + if (flags & ITER_NARROW) + iter->get_scanline = conical_get_scanline_narrow; + else + iter->get_scanline = conical_get_scanline_wide; +} + +PIXMAN_EXPORT pixman_image_t * +pixman_image_create_conical_gradient (pixman_point_fixed_t * center, + pixman_fixed_t angle, + const pixman_gradient_stop_t *stops, + int n_stops) +{ + pixman_image_t *image = _pixman_image_allocate (); + conical_gradient_t *conical; + + if (!image) + return NULL; + + conical = &image->conical; + + if (!_pixman_init_gradient (&conical->common, stops, n_stops)) + { + free (image); + return NULL; + } + + angle = MOD (angle, pixman_int_to_fixed (360)); + + image->type = CONICAL; + + conical->center = *center; + conical->angle = (pixman_fixed_to_double (angle) / 180.0) * M_PI; + + return image; +} + diff --git a/pixman/pixman/pixman-fast-path.c b/pixman/pixman/pixman-fast-path.c index f103b4cf1..868175f60 100644 --- a/pixman/pixman/pixman-fast-path.c +++ b/pixman/pixman/pixman-fast-path.c @@ -188,7 +188,7 @@ fast_composite_in_n_8_8 (pixman_implementation_t *imp, int32_t w; uint16_t t; - src = _pixman_image_get_solid (src_image, dest_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dest_image->bits.format); srca = src >> 24; @@ -312,7 +312,7 @@ fast_composite_over_n_8_8888 (pixman_implementation_t *imp, int dst_stride, mask_stride; int32_t w; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -370,7 +370,7 @@ fast_composite_add_n_8888_8888_ca (pixman_implementation_t *imp, int dst_stride, mask_stride; int32_t w; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -427,7 +427,7 @@ fast_composite_over_n_8888_8888_ca (pixman_implementation_t *imp, int dst_stride, mask_stride; int32_t w; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -494,7 +494,7 @@ fast_composite_over_n_8_0888 (pixman_implementation_t *imp, int dst_stride, mask_stride; int32_t w; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -559,7 +559,7 @@ fast_composite_over_n_8_0565 (pixman_implementation_t *imp, int dst_stride, mask_stride; int32_t w; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -626,7 +626,7 @@ fast_composite_over_n_8888_0565_ca (pixman_implementation_t *imp, int dst_stride, mask_stride; int32_t w; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -1034,7 +1034,7 @@ fast_composite_add_n_8_8 (pixman_implementation_t *imp, PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1); PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); sa = (src >> 24); while (height--) @@ -1146,7 +1146,7 @@ fast_composite_over_n_1_8888 (pixman_implementation_t *imp, if (width <= 0) return; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) return; @@ -1240,7 +1240,7 @@ fast_composite_over_n_1_0565 (pixman_implementation_t *imp, if (width <= 0) return; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) return; @@ -1332,7 +1332,7 @@ fast_composite_solid_fill (pixman_implementation_t *imp, { uint32_t src; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); if (dst_image->bits.format == PIXMAN_a1) { diff --git a/pixman/pixman/pixman-general.c b/pixman/pixman/pixman-general.c index 5a639051b..16ea3a457 100644 --- a/pixman/pixman/pixman-general.c +++ b/pixman/pixman/pixman-general.c @@ -1,315 +1,313 @@ -/*
- * Copyright © 2009 Red Hat, Inc.
- * Copyright © 2000 SuSE, Inc.
- * Copyright © 2007 Red Hat, Inc.
- * Copyright © 2000 Keith Packard, member of The XFree86 Project, Inc.
- * 2005 Lars Knoll & Zack Rusin, Trolltech
- * 2008 Aaron Plattner, NVIDIA Corporation
- *
- * Permission to use, copy, modify, distribute, and sell this software and its
- * documentation for any purpose is hereby granted without fee, provided that
- * the above copyright notice appear in all copies and that both that
- * copyright notice and this permission notice appear in supporting
- * documentation, and that the name of Red Hat not be used in advertising or
- * publicity pertaining to distribution of the software without specific,
- * written prior permission. Red Hat makes no representations about the
- * suitability of this software for any purpose. It is provided "as is"
- * without express or implied warranty.
- *
- * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
- * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
- * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
- * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING
- * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
- * SOFTWARE.
- */
-#ifdef HAVE_CONFIG_H
-#include <config.h>
-#endif
-#include <stdlib.h>
-#include <string.h>
-#include <math.h>
-#include <limits.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include "pixman-private.h"
-#include "pixman-combine32.h"
-#include "pixman-private.h"
-
-#define SCANLINE_BUFFER_LENGTH 8192
-
-static void
-general_composite_rect (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src,
- pixman_image_t * mask,
- pixman_image_t * dest,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint64_t stack_scanline_buffer[(SCANLINE_BUFFER_LENGTH * 3 + 7) / 8];
- uint8_t *scanline_buffer = (uint8_t *) stack_scanline_buffer;
- uint8_t *src_buffer, *mask_buffer, *dest_buffer;
- fetch_scanline_t fetch_src = NULL, fetch_mask = NULL, fetch_dest = NULL;
- pixman_combine_32_func_t compose;
- store_scanline_t store;
- source_image_class_t src_class, mask_class;
- pixman_bool_t component_alpha;
- uint32_t *bits;
- int32_t stride;
- int narrow, Bpp;
- int i;
-
- narrow =
- (src->common.flags & FAST_PATH_NARROW_FORMAT) &&
- (!mask || mask->common.flags & FAST_PATH_NARROW_FORMAT) &&
- (dest->common.flags & FAST_PATH_NARROW_FORMAT);
- Bpp = narrow ? 4 : 8;
-
- if (width * Bpp > SCANLINE_BUFFER_LENGTH)
- {
- scanline_buffer = pixman_malloc_abc (width, 3, Bpp);
-
- if (!scanline_buffer)
- return;
- }
-
- src_buffer = scanline_buffer;
- mask_buffer = src_buffer + width * Bpp;
- dest_buffer = mask_buffer + width * Bpp;
-
- src_class = _pixman_image_classify (src,
- src_x, src_y,
- width, height);
-
- mask_class = SOURCE_IMAGE_CLASS_UNKNOWN;
-
- if (mask)
- {
- mask_class = _pixman_image_classify (mask,
- src_x, src_y,
- width, height);
- }
-
- if (op == PIXMAN_OP_CLEAR)
- fetch_src = NULL;
- else if (narrow)
- fetch_src = _pixman_image_get_scanline_32;
- else
- fetch_src = _pixman_image_get_scanline_64;
-
- if (!mask || op == PIXMAN_OP_CLEAR)
- fetch_mask = NULL;
- else if (narrow)
- fetch_mask = _pixman_image_get_scanline_32;
- else
- fetch_mask = _pixman_image_get_scanline_64;
-
- if (op == PIXMAN_OP_CLEAR || op == PIXMAN_OP_SRC)
- fetch_dest = NULL;
- else if (narrow)
- fetch_dest = _pixman_image_get_scanline_32;
- else
- fetch_dest = _pixman_image_get_scanline_64;
-
- if (narrow)
- store = _pixman_image_store_scanline_32;
- else
- store = _pixman_image_store_scanline_64;
-
- /* Skip the store step and composite directly into the
- * destination if the output format of the compose func matches
- * the destination format.
- *
- * If the destination format is a8r8g8b8 then we can always do
- * this. If it is x8r8g8b8, then we can only do it if the
- * operator doesn't make use of destination alpha.
- */
- if ((dest->bits.format == PIXMAN_a8r8g8b8) ||
- (dest->bits.format == PIXMAN_x8r8g8b8 &&
- (op == PIXMAN_OP_OVER ||
- op == PIXMAN_OP_ADD ||
- op == PIXMAN_OP_SRC ||
- op == PIXMAN_OP_CLEAR ||
- op == PIXMAN_OP_IN_REVERSE ||
- op == PIXMAN_OP_OUT_REVERSE ||
- op == PIXMAN_OP_DST)))
- {
- if (narrow &&
- !dest->common.alpha_map &&
- !dest->bits.write_func)
- {
- store = NULL;
- }
- }
-
- if (!store)
- {
- bits = dest->bits.bits;
- stride = dest->bits.rowstride;
- }
- else
- {
- bits = NULL;
- stride = 0;
- }
-
- component_alpha =
- fetch_src &&
- fetch_mask &&
- mask &&
- mask->common.type == BITS &&
- mask->common.component_alpha &&
- PIXMAN_FORMAT_RGB (mask->bits.format);
-
- if (narrow)
- {
- if (component_alpha)
- compose = _pixman_implementation_combine_32_ca;
- else
- compose = _pixman_implementation_combine_32;
- }
- else
- {
- if (component_alpha)
- compose = (pixman_combine_32_func_t)_pixman_implementation_combine_64_ca;
- else
- compose = (pixman_combine_32_func_t)_pixman_implementation_combine_64;
- }
-
- if (!compose)
- return;
-
- if (!fetch_mask)
- mask_buffer = NULL;
-
- for (i = 0; i < height; ++i)
- {
- /* fill first half of scanline with source */
- if (fetch_src)
- {
- if (fetch_mask)
- {
- /* fetch mask before source so that fetching of
- source can be optimized */
- fetch_mask (mask, mask_x, mask_y + i,
- width, (void *)mask_buffer, 0);
-
- if (mask_class == SOURCE_IMAGE_CLASS_HORIZONTAL)
- fetch_mask = NULL;
- }
-
- if (src_class == SOURCE_IMAGE_CLASS_HORIZONTAL)
- {
- fetch_src (src, src_x, src_y + i,
- width, (void *)src_buffer, 0);
- fetch_src = NULL;
- }
- else
- {
- fetch_src (src, src_x, src_y + i,
- width, (void *)src_buffer, (void *)mask_buffer);
- }
- }
- else if (fetch_mask)
- {
- fetch_mask (mask, mask_x, mask_y + i,
- width, (void *)mask_buffer, 0);
- }
-
- if (store)
- {
- /* fill dest into second half of scanline */
- if (fetch_dest)
- {
- fetch_dest (dest, dest_x, dest_y + i,
- width, (void *)dest_buffer, 0);
- }
-
- /* blend */
- compose (imp->toplevel, op,
- (void *)dest_buffer,
- (void *)src_buffer,
- (void *)mask_buffer,
- width);
-
- /* write back */
- store (&(dest->bits), dest_x, dest_y + i, width,
- (void *)dest_buffer);
- }
- else
- {
- /* blend */
- compose (imp->toplevel, op,
- bits + (dest_y + i) * stride + dest_x,
- (void *)src_buffer, (void *)mask_buffer, width);
- }
- }
-
- if (scanline_buffer != (uint8_t *) stack_scanline_buffer)
- free (scanline_buffer);
-}
-
-static const pixman_fast_path_t general_fast_path[] =
-{
- { PIXMAN_OP_any, PIXMAN_any, 0, PIXMAN_any, 0, PIXMAN_any, 0, general_composite_rect },
- { PIXMAN_OP_NONE }
-};
-
-static pixman_bool_t
-general_blt (pixman_implementation_t *imp,
- uint32_t * src_bits,
- uint32_t * dst_bits,
- int src_stride,
- int dst_stride,
- int src_bpp,
- int dst_bpp,
- int src_x,
- int src_y,
- int dst_x,
- int dst_y,
- int width,
- int height)
-{
- /* We can't blit unless we have sse2 or mmx */
-
- return FALSE;
-}
-
-static pixman_bool_t
-general_fill (pixman_implementation_t *imp,
- uint32_t * bits,
- int stride,
- int bpp,
- int x,
- int y,
- int width,
- int height,
- uint32_t xor)
-{
- return FALSE;
-}
-
-pixman_implementation_t *
-_pixman_implementation_create_general (void)
-{
- pixman_implementation_t *imp = _pixman_implementation_create (NULL, general_fast_path);
-
- _pixman_setup_combiner_functions_32 (imp);
- _pixman_setup_combiner_functions_64 (imp);
-
- imp->blt = general_blt;
- imp->fill = general_fill;
-
- return imp;
-}
-
+/* + * Copyright © 2009 Red Hat, Inc. + * Copyright © 2000 SuSE, Inc. + * Copyright © 2007 Red Hat, Inc. + * Copyright © 2000 Keith Packard, member of The XFree86 Project, Inc. + * 2005 Lars Knoll & Zack Rusin, Trolltech + * 2008 Aaron Plattner, NVIDIA Corporation + * + * Permission to use, copy, modify, distribute, and sell this software and its + * documentation for any purpose is hereby granted without fee, provided that + * the above copyright notice appear in all copies and that both that + * copyright notice and this permission notice appear in supporting + * documentation, and that the name of Red Hat not be used in advertising or + * publicity pertaining to distribution of the software without specific, + * written prior permission. Red Hat makes no representations about the + * suitability of this software for any purpose. It is provided "as is" + * without express or implied warranty. + * + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS + * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND + * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY + * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN + * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING + * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS + * SOFTWARE. + */ +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif +#include <stdlib.h> +#include <string.h> +#include <math.h> +#include <limits.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include "pixman-private.h" +#include "pixman-combine32.h" +#include "pixman-private.h" + +static void +general_src_iter_init (pixman_implementation_t *imp, + pixman_iter_t *iter, + pixman_image_t *image, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags) +{ + iter->image = image; + iter->x = x; + iter->y = y; + iter->width = width; + iter->buffer = (uint32_t *)buffer; + + if (image->type == SOLID) + { + _pixman_solid_fill_iter_init ( + image, iter, x, y, width, height, buffer, flags); + } + else if (image->type == LINEAR) + { + _pixman_linear_gradient_iter_init ( + image, iter, x, y, width, height, buffer, flags); + } + else if (image->type == RADIAL) + { + _pixman_radial_gradient_iter_init ( + image, iter, x, y, width, height, buffer, flags); + } + else if (image->type == CONICAL) + { + _pixman_conical_gradient_iter_init ( + image, iter, x, y, width, height, buffer, flags); + } + else if (image->type == BITS) + { + _pixman_bits_image_src_iter_init ( + image, iter, x, y, width, height, buffer, flags); + } + else + { + _pixman_log_error (FUNC, "Pixman bug: unknown image type\n"); + } +} + +static void +general_dest_iter_init (pixman_implementation_t *imp, + pixman_iter_t *iter, + pixman_image_t *image, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags) +{ + iter->image = image; + iter->x = x; + iter->y = y; + iter->width = width; + iter->buffer = (uint32_t *)buffer; + + if (image->type == BITS) + { + _pixman_bits_image_dest_iter_init ( + image, iter, x, y, width, height, buffer, flags); + } + else + { + _pixman_log_error (FUNC, "Trying to write to a non-writable image"); + } +} + +typedef struct op_info_t op_info_t; +struct op_info_t +{ + uint8_t src, dst; +}; + +#define ITER_IGNORE_BOTH \ + (ITER_IGNORE_ALPHA | ITER_IGNORE_RGB | ITER_LOCALIZED_ALPHA) + +static const op_info_t op_flags[PIXMAN_N_OPERATORS] = +{ + /* Src Dst */ + { ITER_IGNORE_BOTH, ITER_IGNORE_BOTH }, /* CLEAR */ + { ITER_LOCALIZED_ALPHA, ITER_IGNORE_BOTH }, /* SRC */ + { ITER_IGNORE_BOTH, ITER_LOCALIZED_ALPHA }, /* DST */ + { 0, ITER_LOCALIZED_ALPHA }, /* OVER */ + { ITER_LOCALIZED_ALPHA, 0 }, /* OVER_REVERSE */ + { ITER_LOCALIZED_ALPHA, ITER_IGNORE_RGB }, /* IN */ + { ITER_IGNORE_RGB, ITER_LOCALIZED_ALPHA }, /* IN_REVERSE */ + { ITER_LOCALIZED_ALPHA, ITER_IGNORE_RGB }, /* OUT */ + { ITER_IGNORE_RGB, ITER_LOCALIZED_ALPHA }, /* OUT_REVERSE */ + { 0, 0 }, /* ATOP */ + { 0, 0 }, /* ATOP_REVERSE */ + { 0, 0 }, /* XOR */ + { ITER_LOCALIZED_ALPHA, ITER_LOCALIZED_ALPHA }, /* ADD */ + { 0, 0 }, /* SATURATE */ +}; + +#define SCANLINE_BUFFER_LENGTH 8192 + +static void +general_composite_rect (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src, + pixman_image_t * mask, + pixman_image_t * dest, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint64_t stack_scanline_buffer[(SCANLINE_BUFFER_LENGTH * 3 + 7) / 8]; + uint8_t *scanline_buffer = (uint8_t *) stack_scanline_buffer; + uint8_t *src_buffer, *mask_buffer, *dest_buffer; + pixman_iter_t src_iter, mask_iter, dest_iter; + pixman_combine_32_func_t compose; + pixman_bool_t component_alpha; + iter_flags_t narrow, src_flags; + int Bpp; + int i; + + if ((src->common.flags & FAST_PATH_NARROW_FORMAT) && + (!mask || mask->common.flags & FAST_PATH_NARROW_FORMAT) && + (dest->common.flags & FAST_PATH_NARROW_FORMAT)) + { + narrow = ITER_NARROW; + Bpp = 4; + } + else + { + narrow = 0; + Bpp = 8; + } + + if (width * Bpp > SCANLINE_BUFFER_LENGTH) + { + scanline_buffer = pixman_malloc_abc (width, 3, Bpp); + + if (!scanline_buffer) + return; + } + + src_buffer = scanline_buffer; + mask_buffer = src_buffer + width * Bpp; + dest_buffer = mask_buffer + width * Bpp; + + /* src iter */ + src_flags = narrow | op_flags[op].src; + + _pixman_implementation_src_iter_init (imp->toplevel, &src_iter, src, + src_x, src_y, width, height, + src_buffer, src_flags); + + /* mask iter */ + if ((src_flags & (ITER_IGNORE_ALPHA | ITER_IGNORE_RGB)) == + (ITER_IGNORE_ALPHA | ITER_IGNORE_RGB)) + { + /* If it doesn't matter what the source is, then it doesn't matter + * what the mask is + */ + mask = NULL; + } + + component_alpha = + mask && + mask->common.type == BITS && + mask->common.component_alpha && + PIXMAN_FORMAT_RGB (mask->bits.format); + + _pixman_implementation_src_iter_init ( + imp->toplevel, &mask_iter, mask, mask_x, mask_y, width, height, + mask_buffer, narrow | (component_alpha? 0 : ITER_IGNORE_RGB)); + + /* dest iter */ + _pixman_implementation_dest_iter_init (imp->toplevel, &dest_iter, dest, + dest_x, dest_y, width, height, + dest_buffer, + narrow | op_flags[op].dst); + + if (narrow) + { + if (component_alpha) + compose = _pixman_implementation_combine_32_ca; + else + compose = _pixman_implementation_combine_32; + } + else + { + if (component_alpha) + compose = (pixman_combine_32_func_t)_pixman_implementation_combine_64_ca; + else + compose = (pixman_combine_32_func_t)_pixman_implementation_combine_64; + } + + if (!compose) + return; + + for (i = 0; i < height; ++i) + { + uint32_t *s, *m, *d; + + m = mask_iter.get_scanline (&mask_iter, NULL); + s = src_iter.get_scanline (&src_iter, m); + d = dest_iter.get_scanline (&dest_iter, NULL); + + compose (imp->toplevel, op, d, s, m, width); + + dest_iter.write_back (&dest_iter); + } + + if (scanline_buffer != (uint8_t *) stack_scanline_buffer) + free (scanline_buffer); +} + +static const pixman_fast_path_t general_fast_path[] = +{ + { PIXMAN_OP_any, PIXMAN_any, 0, PIXMAN_any, 0, PIXMAN_any, 0, general_composite_rect }, + { PIXMAN_OP_NONE } +}; + +static pixman_bool_t +general_blt (pixman_implementation_t *imp, + uint32_t * src_bits, + uint32_t * dst_bits, + int src_stride, + int dst_stride, + int src_bpp, + int dst_bpp, + int src_x, + int src_y, + int dst_x, + int dst_y, + int width, + int height) +{ + /* We can't blit unless we have sse2 or mmx */ + + return FALSE; +} + +static pixman_bool_t +general_fill (pixman_implementation_t *imp, + uint32_t * bits, + int stride, + int bpp, + int x, + int y, + int width, + int height, + uint32_t xor) +{ + return FALSE; +} + +pixman_implementation_t * +_pixman_implementation_create_general (void) +{ + pixman_implementation_t *imp = _pixman_implementation_create (NULL, general_fast_path); + + _pixman_setup_combiner_functions_32 (imp); + _pixman_setup_combiner_functions_64 (imp); + + imp->blt = general_blt; + imp->fill = general_fill; + imp->src_iter_init = general_src_iter_init; + imp->dest_iter_init = general_dest_iter_init; + + return imp; +} + diff --git a/pixman/pixman/pixman-image.c b/pixman/pixman/pixman-image.c index 48faa3a10..a72299b3c 100644 --- a/pixman/pixman/pixman-image.c +++ b/pixman/pixman/pixman-image.c @@ -1,815 +1,747 @@ -/*
- * Copyright © 2000 SuSE, Inc.
- * Copyright © 2007 Red Hat, Inc.
- *
- * Permission to use, copy, modify, distribute, and sell this software and its
- * documentation for any purpose is hereby granted without fee, provided that
- * the above copyright notice appear in all copies and that both that
- * copyright notice and this permission notice appear in supporting
- * documentation, and that the name of SuSE not be used in advertising or
- * publicity pertaining to distribution of the software without specific,
- * written prior permission. SuSE makes no representations about the
- * suitability of this software for any purpose. It is provided "as is"
- * without express or implied warranty.
- *
- * SuSE DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL SuSE
- * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
- * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
- * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
- */
-
-#ifdef HAVE_CONFIG_H
-#include <config.h>
-#endif
-
-#include <stdlib.h>
-#include <stdio.h>
-#include <string.h>
-#include <assert.h>
-
-#include "pixman-private.h"
-#include "pixman-combine32.h"
-
-pixman_bool_t
-_pixman_init_gradient (gradient_t * gradient,
- const pixman_gradient_stop_t *stops,
- int n_stops)
-{
- return_val_if_fail (n_stops > 0, FALSE);
-
- gradient->stops = pixman_malloc_ab (n_stops, sizeof (pixman_gradient_stop_t));
- if (!gradient->stops)
- return FALSE;
-
- memcpy (gradient->stops, stops, n_stops * sizeof (pixman_gradient_stop_t));
-
- gradient->n_stops = n_stops;
-
- return TRUE;
-}
-
-/*
- * By default, just evaluate the image at 32bpp and expand. Individual image
- * types can plug in a better scanline getter if they want to. For example
- * we could produce smoother gradients by evaluating them at higher color
- * depth, but that's a project for the future.
- */
-void
-_pixman_image_get_scanline_generic_64 (pixman_image_t * image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t * mask)
-{
- uint32_t *mask8 = NULL;
-
- /* Contract the mask image, if one exists, so that the 32-bit fetch
- * function can use it.
- */
- if (mask)
- {
- mask8 = pixman_malloc_ab (width, sizeof(uint32_t));
- if (!mask8)
- return;
-
- pixman_contract (mask8, (uint64_t *)mask, width);
- }
-
- /* Fetch the source image into the first half of buffer. */
- _pixman_image_get_scanline_32 (image, x, y, width, (uint32_t*)buffer, mask8);
-
- /* Expand from 32bpp to 64bpp in place. */
- pixman_expand ((uint64_t *)buffer, buffer, PIXMAN_a8r8g8b8, width);
-
- free (mask8);
-}
-
-pixman_image_t *
-_pixman_image_allocate (void)
-{
- pixman_image_t *image = malloc (sizeof (pixman_image_t));
-
- if (image)
- {
- image_common_t *common = &image->common;
-
- pixman_region32_init (&common->clip_region);
-
- common->alpha_count = 0;
- common->have_clip_region = FALSE;
- common->clip_sources = FALSE;
- common->transform = NULL;
- common->repeat = PIXMAN_REPEAT_NONE;
- common->filter = PIXMAN_FILTER_NEAREST;
- common->filter_params = NULL;
- common->n_filter_params = 0;
- common->alpha_map = NULL;
- common->component_alpha = FALSE;
- common->ref_count = 1;
- common->classify = NULL;
- common->client_clip = FALSE;
- common->destroy_func = NULL;
- common->destroy_data = NULL;
- common->dirty = TRUE;
- }
-
- return image;
-}
-
-source_image_class_t
-_pixman_image_classify (pixman_image_t *image,
- int x,
- int y,
- int width,
- int height)
-{
- if (image->common.classify)
- return image->common.classify (image, x, y, width, height);
- else
- return SOURCE_IMAGE_CLASS_UNKNOWN;
-}
-
-void
-_pixman_image_get_scanline_32 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *mask)
-{
- image->common.get_scanline_32 (image, x, y, width, buffer, mask);
-}
-
-/* Even thought the type of buffer is uint32_t *, the function actually expects
- * a uint64_t *buffer.
- */
-void
-_pixman_image_get_scanline_64 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *unused)
-{
- image->common.get_scanline_64 (image, x, y, width, buffer, unused);
-}
-
-static void
-image_property_changed (pixman_image_t *image)
-{
- image->common.dirty = TRUE;
-}
-
-/* Ref Counting */
-PIXMAN_EXPORT pixman_image_t *
-pixman_image_ref (pixman_image_t *image)
-{
- image->common.ref_count++;
-
- return image;
-}
-
-/* returns TRUE when the image is freed */
-PIXMAN_EXPORT pixman_bool_t
-pixman_image_unref (pixman_image_t *image)
-{
- image_common_t *common = (image_common_t *)image;
-
- common->ref_count--;
-
- if (common->ref_count == 0)
- {
- if (image->common.destroy_func)
- image->common.destroy_func (image, image->common.destroy_data);
-
- pixman_region32_fini (&common->clip_region);
-
- if (common->transform)
- free (common->transform);
-
- if (common->filter_params)
- free (common->filter_params);
-
- if (common->alpha_map)
- pixman_image_unref ((pixman_image_t *)common->alpha_map);
-
- if (image->type == LINEAR ||
- image->type == RADIAL ||
- image->type == CONICAL)
- {
- if (image->gradient.stops)
- free (image->gradient.stops);
- }
-
- if (image->type == BITS && image->bits.free_me)
- free (image->bits.free_me);
-
- free (image);
-
- return TRUE;
- }
-
- return FALSE;
-}
-
-PIXMAN_EXPORT void
-pixman_image_set_destroy_function (pixman_image_t * image,
- pixman_image_destroy_func_t func,
- void * data)
-{
- image->common.destroy_func = func;
- image->common.destroy_data = data;
-}
-
-PIXMAN_EXPORT void *
-pixman_image_get_destroy_data (pixman_image_t *image)
-{
- return image->common.destroy_data;
-}
-
-void
-_pixman_image_reset_clip_region (pixman_image_t *image)
-{
- image->common.have_clip_region = FALSE;
-}
-
-/* Executive Summary: This function is a no-op that only exists
- * for historical reasons.
- *
- * There used to be a bug in the X server where it would rely on
- * out-of-bounds accesses when it was asked to composite with a
- * window as the source. It would create a pixman image pointing
- * to some bogus position in memory, but then set a clip region
- * to the position where the actual bits were.
- *
- * Due to a bug in old versions of pixman, where it would not clip
- * against the image bounds when a clip region was set, this would
- * actually work. So when the pixman bug was fixed, a workaround was
- * added to allow certain out-of-bound accesses. This function disabled
- * those workarounds.
- *
- * Since 0.21.2, pixman doesn't do these workarounds anymore, so now
- * this function is a no-op.
- */
-PIXMAN_EXPORT void
-pixman_disable_out_of_bounds_workaround (void)
-{
-}
-
-static void
-compute_image_info (pixman_image_t *image)
-{
- pixman_format_code_t code;
- uint32_t flags = 0;
-
- /* Transform */
- if (!image->common.transform)
- {
- flags |= (FAST_PATH_ID_TRANSFORM |
- FAST_PATH_X_UNIT_POSITIVE |
- FAST_PATH_Y_UNIT_ZERO |
- FAST_PATH_AFFINE_TRANSFORM);
- }
- else
- {
- flags |= FAST_PATH_HAS_TRANSFORM;
-
- if (image->common.transform->matrix[2][0] == 0 &&
- image->common.transform->matrix[2][1] == 0 &&
- image->common.transform->matrix[2][2] == pixman_fixed_1)
- {
- flags |= FAST_PATH_AFFINE_TRANSFORM;
-
- if (image->common.transform->matrix[0][1] == 0 &&
- image->common.transform->matrix[1][0] == 0)
- {
- flags |= FAST_PATH_SCALE_TRANSFORM;
- }
- }
-
- if (image->common.transform->matrix[0][0] > 0)
- flags |= FAST_PATH_X_UNIT_POSITIVE;
-
- if (image->common.transform->matrix[1][0] == 0)
- flags |= FAST_PATH_Y_UNIT_ZERO;
- }
-
- /* Filter */
- switch (image->common.filter)
- {
- case PIXMAN_FILTER_NEAREST:
- case PIXMAN_FILTER_FAST:
- flags |= (FAST_PATH_NEAREST_FILTER | FAST_PATH_NO_CONVOLUTION_FILTER);
- break;
-
- case PIXMAN_FILTER_BILINEAR:
- case PIXMAN_FILTER_GOOD:
- case PIXMAN_FILTER_BEST:
- flags |= (FAST_PATH_BILINEAR_FILTER | FAST_PATH_NO_CONVOLUTION_FILTER);
- break;
-
- case PIXMAN_FILTER_CONVOLUTION:
- break;
-
- default:
- flags |= FAST_PATH_NO_CONVOLUTION_FILTER;
- break;
- }
-
- /* Repeat mode */
- switch (image->common.repeat)
- {
- case PIXMAN_REPEAT_NONE:
- flags |=
- FAST_PATH_NO_REFLECT_REPEAT |
- FAST_PATH_NO_PAD_REPEAT |
- FAST_PATH_NO_NORMAL_REPEAT;
- break;
-
- case PIXMAN_REPEAT_REFLECT:
- flags |=
- FAST_PATH_NO_PAD_REPEAT |
- FAST_PATH_NO_NONE_REPEAT |
- FAST_PATH_NO_NORMAL_REPEAT;
- break;
-
- case PIXMAN_REPEAT_PAD:
- flags |=
- FAST_PATH_NO_REFLECT_REPEAT |
- FAST_PATH_NO_NONE_REPEAT |
- FAST_PATH_NO_NORMAL_REPEAT;
- break;
-
- default:
- flags |=
- FAST_PATH_NO_REFLECT_REPEAT |
- FAST_PATH_NO_PAD_REPEAT |
- FAST_PATH_NO_NONE_REPEAT;
- break;
- }
-
- /* Component alpha */
- if (image->common.component_alpha)
- flags |= FAST_PATH_COMPONENT_ALPHA;
- else
- flags |= FAST_PATH_UNIFIED_ALPHA;
-
- flags |= (FAST_PATH_NO_ACCESSORS | FAST_PATH_NARROW_FORMAT);
-
- /* Type specific checks */
- switch (image->type)
- {
- case SOLID:
- code = PIXMAN_solid;
-
- if (image->solid.color.alpha == 0xffff)
- flags |= FAST_PATH_IS_OPAQUE;
- break;
-
- case BITS:
- if (image->bits.width == 1 &&
- image->bits.height == 1 &&
- image->common.repeat != PIXMAN_REPEAT_NONE)
- {
- code = PIXMAN_solid;
- }
- else
- {
- code = image->bits.format;
- }
-
- if (!PIXMAN_FORMAT_A (image->bits.format) &&
- PIXMAN_FORMAT_TYPE (image->bits.format) != PIXMAN_TYPE_GRAY &&
- PIXMAN_FORMAT_TYPE (image->bits.format) != PIXMAN_TYPE_COLOR)
- {
- flags |= FAST_PATH_SAMPLES_OPAQUE;
-
- if (image->common.repeat != PIXMAN_REPEAT_NONE)
- flags |= FAST_PATH_IS_OPAQUE;
- }
-
- if (image->bits.read_func || image->bits.write_func)
- flags &= ~FAST_PATH_NO_ACCESSORS;
-
- if (PIXMAN_FORMAT_IS_WIDE (image->bits.format))
- flags &= ~FAST_PATH_NARROW_FORMAT;
- break;
-
- case RADIAL:
- code = PIXMAN_unknown;
-
- /*
- * As explained in pixman-radial-gradient.c, every point of
- * the plane has a valid associated radius (and thus will be
- * colored) if and only if a is negative (i.e. one of the two
- * circles contains the other one).
- */
-
- if (image->radial.a >= 0)
- break;
-
- /* Fall through */
-
- case CONICAL:
- case LINEAR:
- code = PIXMAN_unknown;
-
- if (image->common.repeat != PIXMAN_REPEAT_NONE)
- {
- int i;
-
- flags |= FAST_PATH_IS_OPAQUE;
- for (i = 0; i < image->gradient.n_stops; ++i)
- {
- if (image->gradient.stops[i].color.alpha != 0xffff)
- {
- flags &= ~FAST_PATH_IS_OPAQUE;
- break;
- }
- }
- }
- break;
-
- default:
- code = PIXMAN_unknown;
- break;
- }
-
- /* Alpha map */
- if (!image->common.alpha_map)
- {
- flags |= FAST_PATH_NO_ALPHA_MAP;
- }
- else
- {
- if (PIXMAN_FORMAT_IS_WIDE (image->common.alpha_map->format))
- flags &= ~FAST_PATH_NARROW_FORMAT;
- }
-
- /* Both alpha maps and convolution filters can introduce
- * non-opaqueness in otherwise opaque images. Also
- * an image with component alpha turned on is only opaque
- * if all channels are opaque, so we simply turn it off
- * unconditionally for those images.
- */
- if (image->common.alpha_map ||
- image->common.filter == PIXMAN_FILTER_CONVOLUTION ||
- image->common.component_alpha)
- {
- flags &= ~(FAST_PATH_IS_OPAQUE | FAST_PATH_SAMPLES_OPAQUE);
- }
-
- image->common.flags = flags;
- image->common.extended_format_code = code;
-}
-
-void
-_pixman_image_validate (pixman_image_t *image)
-{
- if (image->common.dirty)
- {
- compute_image_info (image);
-
- /* It is important that property_changed is
- * called *after* compute_image_info() because
- * property_changed() can make use of the flags
- * to set up accessors etc.
- */
- image->common.property_changed (image);
-
- image->common.dirty = FALSE;
- }
-
- if (image->common.alpha_map)
- _pixman_image_validate ((pixman_image_t *)image->common.alpha_map);
-}
-
-PIXMAN_EXPORT pixman_bool_t
-pixman_image_set_clip_region32 (pixman_image_t * image,
- pixman_region32_t *region)
-{
- image_common_t *common = (image_common_t *)image;
- pixman_bool_t result;
-
- if (region)
- {
- if ((result = pixman_region32_copy (&common->clip_region, region)))
- image->common.have_clip_region = TRUE;
- }
- else
- {
- _pixman_image_reset_clip_region (image);
-
- result = TRUE;
- }
-
- image_property_changed (image);
-
- return result;
-}
-
-PIXMAN_EXPORT pixman_bool_t
-pixman_image_set_clip_region (pixman_image_t * image,
- pixman_region16_t *region)
-{
- image_common_t *common = (image_common_t *)image;
- pixman_bool_t result;
-
- if (region)
- {
- if ((result = pixman_region32_copy_from_region16 (&common->clip_region, region)))
- image->common.have_clip_region = TRUE;
- }
- else
- {
- _pixman_image_reset_clip_region (image);
-
- result = TRUE;
- }
-
- image_property_changed (image);
-
- return result;
-}
-
-PIXMAN_EXPORT void
-pixman_image_set_has_client_clip (pixman_image_t *image,
- pixman_bool_t client_clip)
-{
- image->common.client_clip = client_clip;
-}
-
-PIXMAN_EXPORT pixman_bool_t
-pixman_image_set_transform (pixman_image_t * image,
- const pixman_transform_t *transform)
-{
- static const pixman_transform_t id =
- {
- { { pixman_fixed_1, 0, 0 },
- { 0, pixman_fixed_1, 0 },
- { 0, 0, pixman_fixed_1 } }
- };
-
- image_common_t *common = (image_common_t *)image;
- pixman_bool_t result;
-
- if (common->transform == transform)
- return TRUE;
-
- if (memcmp (&id, transform, sizeof (pixman_transform_t)) == 0)
- {
- free (common->transform);
- common->transform = NULL;
- result = TRUE;
-
- goto out;
- }
-
- if (common->transform == NULL)
- common->transform = malloc (sizeof (pixman_transform_t));
-
- if (common->transform == NULL)
- {
- result = FALSE;
-
- goto out;
- }
-
- memcpy (common->transform, transform, sizeof(pixman_transform_t));
-
- result = TRUE;
-
-out:
- image_property_changed (image);
-
- return result;
-}
-
-PIXMAN_EXPORT void
-pixman_image_set_repeat (pixman_image_t *image,
- pixman_repeat_t repeat)
-{
- image->common.repeat = repeat;
-
- image_property_changed (image);
-}
-
-PIXMAN_EXPORT pixman_bool_t
-pixman_image_set_filter (pixman_image_t * image,
- pixman_filter_t filter,
- const pixman_fixed_t *params,
- int n_params)
-{
- image_common_t *common = (image_common_t *)image;
- pixman_fixed_t *new_params;
-
- if (params == common->filter_params && filter == common->filter)
- return TRUE;
-
- new_params = NULL;
- if (params)
- {
- new_params = pixman_malloc_ab (n_params, sizeof (pixman_fixed_t));
- if (!new_params)
- return FALSE;
-
- memcpy (new_params,
- params, n_params * sizeof (pixman_fixed_t));
- }
-
- common->filter = filter;
-
- if (common->filter_params)
- free (common->filter_params);
-
- common->filter_params = new_params;
- common->n_filter_params = n_params;
-
- image_property_changed (image);
- return TRUE;
-}
-
-PIXMAN_EXPORT void
-pixman_image_set_source_clipping (pixman_image_t *image,
- pixman_bool_t clip_sources)
-{
- image->common.clip_sources = clip_sources;
-
- image_property_changed (image);
-}
-
-/* Unlike all the other property setters, this function does not
- * copy the content of indexed. Doing this copying is simply
- * way, way too expensive.
- */
-PIXMAN_EXPORT void
-pixman_image_set_indexed (pixman_image_t * image,
- const pixman_indexed_t *indexed)
-{
- bits_image_t *bits = (bits_image_t *)image;
-
- bits->indexed = indexed;
-
- image_property_changed (image);
-}
-
-PIXMAN_EXPORT void
-pixman_image_set_alpha_map (pixman_image_t *image,
- pixman_image_t *alpha_map,
- int16_t x,
- int16_t y)
-{
- image_common_t *common = (image_common_t *)image;
-
- return_if_fail (!alpha_map || alpha_map->type == BITS);
-
- if (alpha_map && common->alpha_count > 0)
- {
- /* If this image is being used as an alpha map itself,
- * then you can't give it an alpha map of its own.
- */
- return;
- }
-
- if (alpha_map && alpha_map->common.alpha_map)
- {
- /* If the image has an alpha map of its own,
- * then it can't be used as an alpha map itself
- */
- return;
- }
-
- if (common->alpha_map != (bits_image_t *)alpha_map)
- {
- if (common->alpha_map)
- {
- common->alpha_map->common.alpha_count--;
-
- pixman_image_unref ((pixman_image_t *)common->alpha_map);
- }
-
- if (alpha_map)
- {
- common->alpha_map = (bits_image_t *)pixman_image_ref (alpha_map);
-
- common->alpha_map->common.alpha_count++;
- }
- else
- {
- common->alpha_map = NULL;
- }
- }
-
- common->alpha_origin_x = x;
- common->alpha_origin_y = y;
-
- image_property_changed (image);
-}
-
-PIXMAN_EXPORT void
-pixman_image_set_component_alpha (pixman_image_t *image,
- pixman_bool_t component_alpha)
-{
- image->common.component_alpha = component_alpha;
-
- image_property_changed (image);
-}
-
-PIXMAN_EXPORT pixman_bool_t
-pixman_image_get_component_alpha (pixman_image_t *image)
-{
- return image->common.component_alpha;
-}
-
-PIXMAN_EXPORT void
-pixman_image_set_accessors (pixman_image_t * image,
- pixman_read_memory_func_t read_func,
- pixman_write_memory_func_t write_func)
-{
- return_if_fail (image != NULL);
-
- if (image->type == BITS)
- {
- image->bits.read_func = read_func;
- image->bits.write_func = write_func;
-
- image_property_changed (image);
- }
-}
-
-PIXMAN_EXPORT uint32_t *
-pixman_image_get_data (pixman_image_t *image)
-{
- if (image->type == BITS)
- return image->bits.bits;
-
- return NULL;
-}
-
-PIXMAN_EXPORT int
-pixman_image_get_width (pixman_image_t *image)
-{
- if (image->type == BITS)
- return image->bits.width;
-
- return 0;
-}
-
-PIXMAN_EXPORT int
-pixman_image_get_height (pixman_image_t *image)
-{
- if (image->type == BITS)
- return image->bits.height;
-
- return 0;
-}
-
-PIXMAN_EXPORT int
-pixman_image_get_stride (pixman_image_t *image)
-{
- if (image->type == BITS)
- return image->bits.rowstride * (int) sizeof (uint32_t);
-
- return 0;
-}
-
-PIXMAN_EXPORT int
-pixman_image_get_depth (pixman_image_t *image)
-{
- if (image->type == BITS)
- return PIXMAN_FORMAT_DEPTH (image->bits.format);
-
- return 0;
-}
-
-PIXMAN_EXPORT pixman_format_code_t
-pixman_image_get_format (pixman_image_t *image)
-{
- if (image->type == BITS)
- return image->bits.format;
-
- return 0;
-}
-
-uint32_t
-_pixman_image_get_solid (pixman_image_t * image,
- pixman_format_code_t format)
-{
- uint32_t result;
-
- _pixman_image_get_scanline_32 (image, 0, 0, 1, &result, NULL);
-
- /* If necessary, convert RGB <--> BGR. */
- if (PIXMAN_FORMAT_TYPE (format) != PIXMAN_TYPE_ARGB)
- {
- result = (((result & 0xff000000) >> 0) |
- ((result & 0x00ff0000) >> 16) |
- ((result & 0x0000ff00) >> 0) |
- ((result & 0x000000ff) << 16));
- }
-
- return result;
-}
+/* + * Copyright © 2000 SuSE, Inc. + * Copyright © 2007 Red Hat, Inc. + * + * Permission to use, copy, modify, distribute, and sell this software and its + * documentation for any purpose is hereby granted without fee, provided that + * the above copyright notice appear in all copies and that both that + * copyright notice and this permission notice appear in supporting + * documentation, and that the name of SuSE not be used in advertising or + * publicity pertaining to distribution of the software without specific, + * written prior permission. SuSE makes no representations about the + * suitability of this software for any purpose. It is provided "as is" + * without express or implied warranty. + * + * SuSE DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL SuSE + * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION + * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN + * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. + */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#include <stdlib.h> +#include <stdio.h> +#include <string.h> +#include <assert.h> + +#include "pixman-private.h" +#include "pixman-combine32.h" + +pixman_bool_t +_pixman_init_gradient (gradient_t * gradient, + const pixman_gradient_stop_t *stops, + int n_stops) +{ + return_val_if_fail (n_stops > 0, FALSE); + + gradient->stops = pixman_malloc_ab (n_stops, sizeof (pixman_gradient_stop_t)); + if (!gradient->stops) + return FALSE; + + memcpy (gradient->stops, stops, n_stops * sizeof (pixman_gradient_stop_t)); + + gradient->n_stops = n_stops; + + return TRUE; +} + +pixman_image_t * +_pixman_image_allocate (void) +{ + pixman_image_t *image = malloc (sizeof (pixman_image_t)); + + if (image) + { + image_common_t *common = &image->common; + + pixman_region32_init (&common->clip_region); + + common->alpha_count = 0; + common->have_clip_region = FALSE; + common->clip_sources = FALSE; + common->transform = NULL; + common->repeat = PIXMAN_REPEAT_NONE; + common->filter = PIXMAN_FILTER_NEAREST; + common->filter_params = NULL; + common->n_filter_params = 0; + common->alpha_map = NULL; + common->component_alpha = FALSE; + common->ref_count = 1; + common->property_changed = NULL; + common->client_clip = FALSE; + common->destroy_func = NULL; + common->destroy_data = NULL; + common->dirty = TRUE; + } + + return image; +} + +static void +image_property_changed (pixman_image_t *image) +{ + image->common.dirty = TRUE; +} + +/* Ref Counting */ +PIXMAN_EXPORT pixman_image_t * +pixman_image_ref (pixman_image_t *image) +{ + image->common.ref_count++; + + return image; +} + +/* returns TRUE when the image is freed */ +PIXMAN_EXPORT pixman_bool_t +pixman_image_unref (pixman_image_t *image) +{ + image_common_t *common = (image_common_t *)image; + + common->ref_count--; + + if (common->ref_count == 0) + { + if (image->common.destroy_func) + image->common.destroy_func (image, image->common.destroy_data); + + pixman_region32_fini (&common->clip_region); + + if (common->transform) + free (common->transform); + + if (common->filter_params) + free (common->filter_params); + + if (common->alpha_map) + pixman_image_unref ((pixman_image_t *)common->alpha_map); + + if (image->type == LINEAR || + image->type == RADIAL || + image->type == CONICAL) + { + if (image->gradient.stops) + free (image->gradient.stops); + } + + if (image->type == BITS && image->bits.free_me) + free (image->bits.free_me); + + free (image); + + return TRUE; + } + + return FALSE; +} + +PIXMAN_EXPORT void +pixman_image_set_destroy_function (pixman_image_t * image, + pixman_image_destroy_func_t func, + void * data) +{ + image->common.destroy_func = func; + image->common.destroy_data = data; +} + +PIXMAN_EXPORT void * +pixman_image_get_destroy_data (pixman_image_t *image) +{ + return image->common.destroy_data; +} + +void +_pixman_image_reset_clip_region (pixman_image_t *image) +{ + image->common.have_clip_region = FALSE; +} + +/* Executive Summary: This function is a no-op that only exists + * for historical reasons. + * + * There used to be a bug in the X server where it would rely on + * out-of-bounds accesses when it was asked to composite with a + * window as the source. It would create a pixman image pointing + * to some bogus position in memory, but then set a clip region + * to the position where the actual bits were. + * + * Due to a bug in old versions of pixman, where it would not clip + * against the image bounds when a clip region was set, this would + * actually work. So when the pixman bug was fixed, a workaround was + * added to allow certain out-of-bound accesses. This function disabled + * those workarounds. + * + * Since 0.21.2, pixman doesn't do these workarounds anymore, so now + * this function is a no-op. + */ +PIXMAN_EXPORT void +pixman_disable_out_of_bounds_workaround (void) +{ +} + +static void +compute_image_info (pixman_image_t *image) +{ + pixman_format_code_t code; + uint32_t flags = 0; + + /* Transform */ + if (!image->common.transform) + { + flags |= (FAST_PATH_ID_TRANSFORM | + FAST_PATH_X_UNIT_POSITIVE | + FAST_PATH_Y_UNIT_ZERO | + FAST_PATH_AFFINE_TRANSFORM); + } + else + { + flags |= FAST_PATH_HAS_TRANSFORM; + + if (image->common.transform->matrix[2][0] == 0 && + image->common.transform->matrix[2][1] == 0 && + image->common.transform->matrix[2][2] == pixman_fixed_1) + { + flags |= FAST_PATH_AFFINE_TRANSFORM; + + if (image->common.transform->matrix[0][1] == 0 && + image->common.transform->matrix[1][0] == 0) + { + flags |= FAST_PATH_SCALE_TRANSFORM; + } + } + + if (image->common.transform->matrix[0][0] > 0) + flags |= FAST_PATH_X_UNIT_POSITIVE; + + if (image->common.transform->matrix[1][0] == 0) + flags |= FAST_PATH_Y_UNIT_ZERO; + } + + /* Filter */ + switch (image->common.filter) + { + case PIXMAN_FILTER_NEAREST: + case PIXMAN_FILTER_FAST: + flags |= (FAST_PATH_NEAREST_FILTER | FAST_PATH_NO_CONVOLUTION_FILTER); + break; + + case PIXMAN_FILTER_BILINEAR: + case PIXMAN_FILTER_GOOD: + case PIXMAN_FILTER_BEST: + flags |= (FAST_PATH_BILINEAR_FILTER | FAST_PATH_NO_CONVOLUTION_FILTER); + break; + + case PIXMAN_FILTER_CONVOLUTION: + break; + + default: + flags |= FAST_PATH_NO_CONVOLUTION_FILTER; + break; + } + + /* Repeat mode */ + switch (image->common.repeat) + { + case PIXMAN_REPEAT_NONE: + flags |= + FAST_PATH_NO_REFLECT_REPEAT | + FAST_PATH_NO_PAD_REPEAT | + FAST_PATH_NO_NORMAL_REPEAT; + break; + + case PIXMAN_REPEAT_REFLECT: + flags |= + FAST_PATH_NO_PAD_REPEAT | + FAST_PATH_NO_NONE_REPEAT | + FAST_PATH_NO_NORMAL_REPEAT; + break; + + case PIXMAN_REPEAT_PAD: + flags |= + FAST_PATH_NO_REFLECT_REPEAT | + FAST_PATH_NO_NONE_REPEAT | + FAST_PATH_NO_NORMAL_REPEAT; + break; + + default: + flags |= + FAST_PATH_NO_REFLECT_REPEAT | + FAST_PATH_NO_PAD_REPEAT | + FAST_PATH_NO_NONE_REPEAT; + break; + } + + /* Component alpha */ + if (image->common.component_alpha) + flags |= FAST_PATH_COMPONENT_ALPHA; + else + flags |= FAST_PATH_UNIFIED_ALPHA; + + flags |= (FAST_PATH_NO_ACCESSORS | FAST_PATH_NARROW_FORMAT); + + /* Type specific checks */ + switch (image->type) + { + case SOLID: + code = PIXMAN_solid; + + if (image->solid.color.alpha == 0xffff) + flags |= FAST_PATH_IS_OPAQUE; + break; + + case BITS: + if (image->bits.width == 1 && + image->bits.height == 1 && + image->common.repeat != PIXMAN_REPEAT_NONE) + { + code = PIXMAN_solid; + } + else + { + code = image->bits.format; + } + + if (!PIXMAN_FORMAT_A (image->bits.format) && + PIXMAN_FORMAT_TYPE (image->bits.format) != PIXMAN_TYPE_GRAY && + PIXMAN_FORMAT_TYPE (image->bits.format) != PIXMAN_TYPE_COLOR) + { + flags |= FAST_PATH_SAMPLES_OPAQUE; + + if (image->common.repeat != PIXMAN_REPEAT_NONE) + flags |= FAST_PATH_IS_OPAQUE; + } + + if (image->bits.read_func || image->bits.write_func) + flags &= ~FAST_PATH_NO_ACCESSORS; + + if (PIXMAN_FORMAT_IS_WIDE (image->bits.format)) + flags &= ~FAST_PATH_NARROW_FORMAT; + break; + + case RADIAL: + code = PIXMAN_unknown; + + /* + * As explained in pixman-radial-gradient.c, every point of + * the plane has a valid associated radius (and thus will be + * colored) if and only if a is negative (i.e. one of the two + * circles contains the other one). + */ + + if (image->radial.a >= 0) + break; + + /* Fall through */ + + case CONICAL: + case LINEAR: + code = PIXMAN_unknown; + + if (image->common.repeat != PIXMAN_REPEAT_NONE) + { + int i; + + flags |= FAST_PATH_IS_OPAQUE; + for (i = 0; i < image->gradient.n_stops; ++i) + { + if (image->gradient.stops[i].color.alpha != 0xffff) + { + flags &= ~FAST_PATH_IS_OPAQUE; + break; + } + } + } + break; + + default: + code = PIXMAN_unknown; + break; + } + + /* Alpha map */ + if (!image->common.alpha_map) + { + flags |= FAST_PATH_NO_ALPHA_MAP; + } + else + { + if (PIXMAN_FORMAT_IS_WIDE (image->common.alpha_map->format)) + flags &= ~FAST_PATH_NARROW_FORMAT; + } + + /* Both alpha maps and convolution filters can introduce + * non-opaqueness in otherwise opaque images. Also + * an image with component alpha turned on is only opaque + * if all channels are opaque, so we simply turn it off + * unconditionally for those images. + */ + if (image->common.alpha_map || + image->common.filter == PIXMAN_FILTER_CONVOLUTION || + image->common.component_alpha) + { + flags &= ~(FAST_PATH_IS_OPAQUE | FAST_PATH_SAMPLES_OPAQUE); + } + + image->common.flags = flags; + image->common.extended_format_code = code; +} + +void +_pixman_image_validate (pixman_image_t *image) +{ + if (image->common.dirty) + { + compute_image_info (image); + + /* It is important that property_changed is + * called *after* compute_image_info() because + * property_changed() can make use of the flags + * to set up accessors etc. + */ + if (image->common.property_changed) + image->common.property_changed (image); + + image->common.dirty = FALSE; + } + + if (image->common.alpha_map) + _pixman_image_validate ((pixman_image_t *)image->common.alpha_map); +} + +PIXMAN_EXPORT pixman_bool_t +pixman_image_set_clip_region32 (pixman_image_t * image, + pixman_region32_t *region) +{ + image_common_t *common = (image_common_t *)image; + pixman_bool_t result; + + if (region) + { + if ((result = pixman_region32_copy (&common->clip_region, region))) + image->common.have_clip_region = TRUE; + } + else + { + _pixman_image_reset_clip_region (image); + + result = TRUE; + } + + image_property_changed (image); + + return result; +} + +PIXMAN_EXPORT pixman_bool_t +pixman_image_set_clip_region (pixman_image_t * image, + pixman_region16_t *region) +{ + image_common_t *common = (image_common_t *)image; + pixman_bool_t result; + + if (region) + { + if ((result = pixman_region32_copy_from_region16 (&common->clip_region, region))) + image->common.have_clip_region = TRUE; + } + else + { + _pixman_image_reset_clip_region (image); + + result = TRUE; + } + + image_property_changed (image); + + return result; +} + +PIXMAN_EXPORT void +pixman_image_set_has_client_clip (pixman_image_t *image, + pixman_bool_t client_clip) +{ + image->common.client_clip = client_clip; +} + +PIXMAN_EXPORT pixman_bool_t +pixman_image_set_transform (pixman_image_t * image, + const pixman_transform_t *transform) +{ + static const pixman_transform_t id = + { + { { pixman_fixed_1, 0, 0 }, + { 0, pixman_fixed_1, 0 }, + { 0, 0, pixman_fixed_1 } } + }; + + image_common_t *common = (image_common_t *)image; + pixman_bool_t result; + + if (common->transform == transform) + return TRUE; + + if (memcmp (&id, transform, sizeof (pixman_transform_t)) == 0) + { + free (common->transform); + common->transform = NULL; + result = TRUE; + + goto out; + } + + if (common->transform == NULL) + common->transform = malloc (sizeof (pixman_transform_t)); + + if (common->transform == NULL) + { + result = FALSE; + + goto out; + } + + memcpy (common->transform, transform, sizeof(pixman_transform_t)); + + result = TRUE; + +out: + image_property_changed (image); + + return result; +} + +PIXMAN_EXPORT void +pixman_image_set_repeat (pixman_image_t *image, + pixman_repeat_t repeat) +{ + image->common.repeat = repeat; + + image_property_changed (image); +} + +PIXMAN_EXPORT pixman_bool_t +pixman_image_set_filter (pixman_image_t * image, + pixman_filter_t filter, + const pixman_fixed_t *params, + int n_params) +{ + image_common_t *common = (image_common_t *)image; + pixman_fixed_t *new_params; + + if (params == common->filter_params && filter == common->filter) + return TRUE; + + new_params = NULL; + if (params) + { + new_params = pixman_malloc_ab (n_params, sizeof (pixman_fixed_t)); + if (!new_params) + return FALSE; + + memcpy (new_params, + params, n_params * sizeof (pixman_fixed_t)); + } + + common->filter = filter; + + if (common->filter_params) + free (common->filter_params); + + common->filter_params = new_params; + common->n_filter_params = n_params; + + image_property_changed (image); + return TRUE; +} + +PIXMAN_EXPORT void +pixman_image_set_source_clipping (pixman_image_t *image, + pixman_bool_t clip_sources) +{ + image->common.clip_sources = clip_sources; + + image_property_changed (image); +} + +/* Unlike all the other property setters, this function does not + * copy the content of indexed. Doing this copying is simply + * way, way too expensive. + */ +PIXMAN_EXPORT void +pixman_image_set_indexed (pixman_image_t * image, + const pixman_indexed_t *indexed) +{ + bits_image_t *bits = (bits_image_t *)image; + + bits->indexed = indexed; + + image_property_changed (image); +} + +PIXMAN_EXPORT void +pixman_image_set_alpha_map (pixman_image_t *image, + pixman_image_t *alpha_map, + int16_t x, + int16_t y) +{ + image_common_t *common = (image_common_t *)image; + + return_if_fail (!alpha_map || alpha_map->type == BITS); + + if (alpha_map && common->alpha_count > 0) + { + /* If this image is being used as an alpha map itself, + * then you can't give it an alpha map of its own. + */ + return; + } + + if (alpha_map && alpha_map->common.alpha_map) + { + /* If the image has an alpha map of its own, + * then it can't be used as an alpha map itself + */ + return; + } + + if (common->alpha_map != (bits_image_t *)alpha_map) + { + if (common->alpha_map) + { + common->alpha_map->common.alpha_count--; + + pixman_image_unref ((pixman_image_t *)common->alpha_map); + } + + if (alpha_map) + { + common->alpha_map = (bits_image_t *)pixman_image_ref (alpha_map); + + common->alpha_map->common.alpha_count++; + } + else + { + common->alpha_map = NULL; + } + } + + common->alpha_origin_x = x; + common->alpha_origin_y = y; + + image_property_changed (image); +} + +PIXMAN_EXPORT void +pixman_image_set_component_alpha (pixman_image_t *image, + pixman_bool_t component_alpha) +{ + image->common.component_alpha = component_alpha; + + image_property_changed (image); +} + +PIXMAN_EXPORT pixman_bool_t +pixman_image_get_component_alpha (pixman_image_t *image) +{ + return image->common.component_alpha; +} + +PIXMAN_EXPORT void +pixman_image_set_accessors (pixman_image_t * image, + pixman_read_memory_func_t read_func, + pixman_write_memory_func_t write_func) +{ + return_if_fail (image != NULL); + + if (image->type == BITS) + { + image->bits.read_func = read_func; + image->bits.write_func = write_func; + + image_property_changed (image); + } +} + +PIXMAN_EXPORT uint32_t * +pixman_image_get_data (pixman_image_t *image) +{ + if (image->type == BITS) + return image->bits.bits; + + return NULL; +} + +PIXMAN_EXPORT int +pixman_image_get_width (pixman_image_t *image) +{ + if (image->type == BITS) + return image->bits.width; + + return 0; +} + +PIXMAN_EXPORT int +pixman_image_get_height (pixman_image_t *image) +{ + if (image->type == BITS) + return image->bits.height; + + return 0; +} + +PIXMAN_EXPORT int +pixman_image_get_stride (pixman_image_t *image) +{ + if (image->type == BITS) + return image->bits.rowstride * (int) sizeof (uint32_t); + + return 0; +} + +PIXMAN_EXPORT int +pixman_image_get_depth (pixman_image_t *image) +{ + if (image->type == BITS) + return PIXMAN_FORMAT_DEPTH (image->bits.format); + + return 0; +} + +PIXMAN_EXPORT pixman_format_code_t +pixman_image_get_format (pixman_image_t *image) +{ + if (image->type == BITS) + return image->bits.format; + + return 0; +} + +uint32_t +_pixman_image_get_solid (pixman_implementation_t *imp, + pixman_image_t * image, + pixman_format_code_t format) +{ + uint32_t result; + pixman_iter_t iter; + + _pixman_implementation_src_iter_init ( + imp, &iter, image, 0, 0, 1, 1, + (uint8_t *)&result, ITER_NARROW); + + result = *iter.get_scanline (&iter, NULL); + + /* If necessary, convert RGB <--> BGR. */ + if (PIXMAN_FORMAT_TYPE (format) != PIXMAN_TYPE_ARGB) + { + result = (((result & 0xff000000) >> 0) | + ((result & 0x00ff0000) >> 16) | + ((result & 0x0000ff00) >> 0) | + ((result & 0x000000ff) << 16)); + } + + return result; +} diff --git a/pixman/pixman/pixman-implementation.c b/pixman/pixman/pixman-implementation.c index bc3749ef5..adaf9c61e 100644 --- a/pixman/pixman/pixman-implementation.c +++ b/pixman/pixman/pixman-implementation.c @@ -111,6 +111,36 @@ delegate_fill (pixman_implementation_t *imp, imp->delegate, bits, stride, bpp, x, y, width, height, xor); } +static void +delegate_src_iter_init (pixman_implementation_t *imp, + pixman_iter_t * iter, + pixman_image_t * image, + int x, + int y, + int width, + int height, + uint8_t * buffer, + iter_flags_t flags) +{ + _pixman_implementation_src_iter_init ( + imp->delegate, iter, image, x, y, width, height, buffer, flags); +} + +static void +delegate_dest_iter_init (pixman_implementation_t *imp, + pixman_iter_t * iter, + pixman_image_t * image, + int x, + int y, + int width, + int height, + uint8_t * buffer, + iter_flags_t flags) +{ + _pixman_implementation_dest_iter_init ( + imp->delegate, iter, image, x, y, width, height, buffer, flags); +} + pixman_implementation_t * _pixman_implementation_create (pixman_implementation_t *delegate, const pixman_fast_path_t *fast_paths) @@ -133,6 +163,8 @@ _pixman_implementation_create (pixman_implementation_t *delegate, */ imp->blt = delegate_blt; imp->fill = delegate_fill; + imp->src_iter_init = delegate_src_iter_init; + imp->dest_iter_init = delegate_dest_iter_init; for (i = 0; i < PIXMAN_N_OPERATORS; ++i) { @@ -143,7 +175,7 @@ _pixman_implementation_create (pixman_implementation_t *delegate, } imp->fast_paths = fast_paths; - + return imp; } @@ -225,3 +257,50 @@ _pixman_implementation_fill (pixman_implementation_t *imp, return (*imp->fill) (imp, bits, stride, bpp, x, y, width, height, xor); } +static uint32_t * +get_scanline_null (pixman_iter_t *iter, const uint32_t *mask) +{ + return NULL; +} + +void +_pixman_implementation_src_iter_init (pixman_implementation_t *imp, + pixman_iter_t *iter, + pixman_image_t *image, + int x, + int y, + int width, + int height, + uint8_t *buffer, + iter_flags_t flags) +{ + if (!image) + { + iter->get_scanline = get_scanline_null; + } + else if ((flags & (ITER_IGNORE_ALPHA | ITER_IGNORE_RGB)) == + (ITER_IGNORE_ALPHA | ITER_IGNORE_RGB)) + { + iter->get_scanline = _pixman_iter_get_scanline_noop; + } + else + { + (*imp->src_iter_init) ( + imp, iter, image, x, y, width, height, buffer, flags); + } +} + +void +_pixman_implementation_dest_iter_init (pixman_implementation_t *imp, + pixman_iter_t *iter, + pixman_image_t *image, + int x, + int y, + int width, + int height, + uint8_t *buffer, + iter_flags_t flags) +{ + (*imp->dest_iter_init) ( + imp, iter, image, x, y, width, height, buffer, flags); +} diff --git a/pixman/pixman/pixman-linear-gradient.c b/pixman/pixman/pixman-linear-gradient.c index 1756b4a0e..07303fc03 100644 --- a/pixman/pixman/pixman-linear-gradient.c +++ b/pixman/pixman/pixman-linear-gradient.c @@ -1,260 +1,292 @@ -/* -*- Mode: c; c-basic-offset: 4; tab-width: 8; indent-tabs-mode: t; -*- */
-/*
- * Copyright © 2000 SuSE, Inc.
- * Copyright © 2007 Red Hat, Inc.
- * Copyright © 2000 Keith Packard, member of The XFree86 Project, Inc.
- * 2005 Lars Knoll & Zack Rusin, Trolltech
- *
- * Permission to use, copy, modify, distribute, and sell this software and its
- * documentation for any purpose is hereby granted without fee, provided that
- * the above copyright notice appear in all copies and that both that
- * copyright notice and this permission notice appear in supporting
- * documentation, and that the name of Keith Packard not be used in
- * advertising or publicity pertaining to distribution of the software without
- * specific, written prior permission. Keith Packard makes no
- * representations about the suitability of this software for any purpose. It
- * is provided "as is" without express or implied warranty.
- *
- * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
- * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
- * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
- * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING
- * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
- * SOFTWARE.
- */
-
-#ifdef HAVE_CONFIG_H
-#include <config.h>
-#endif
-#include <stdlib.h>
-#include "pixman-private.h"
-
-static source_image_class_t
-linear_gradient_classify (pixman_image_t *image,
- int x,
- int y,
- int width,
- int height)
-{
- linear_gradient_t *linear = (linear_gradient_t *)image;
- pixman_vector_t v;
- pixman_fixed_32_32_t l;
- pixman_fixed_48_16_t dx, dy;
- double inc;
- source_image_class_t class;
-
- class = SOURCE_IMAGE_CLASS_UNKNOWN;
-
- if (image->common.transform)
- {
- /* projective transformation */
- if (image->common.transform->matrix[2][0] != 0 ||
- image->common.transform->matrix[2][1] != 0 ||
- image->common.transform->matrix[2][2] == 0)
- {
- return class;
- }
-
- v.vector[0] = image->common.transform->matrix[0][1];
- v.vector[1] = image->common.transform->matrix[1][1];
- v.vector[2] = image->common.transform->matrix[2][2];
- }
- else
- {
- v.vector[0] = 0;
- v.vector[1] = pixman_fixed_1;
- v.vector[2] = pixman_fixed_1;
- }
-
- dx = linear->p2.x - linear->p1.x;
- dy = linear->p2.y - linear->p1.y;
-
- l = dx * dx + dy * dy;
-
- if (l == 0)
- return class;
-
- /*
- * compute how much the input of the gradient walked changes
- * when moving vertically through the whole image
- */
- inc = height * (double) pixman_fixed_1 * pixman_fixed_1 *
- (dx * v.vector[0] + dy * v.vector[1]) /
- (v.vector[2] * (double) l);
-
- /* check that casting to integer would result in 0 */
- if (-1 < inc && inc < 1)
- class = SOURCE_IMAGE_CLASS_HORIZONTAL;
-
- return class;
-}
-
-static void
-linear_gradient_get_scanline_32 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *mask)
-{
- pixman_vector_t v, unit;
- pixman_fixed_32_32_t l;
- pixman_fixed_48_16_t dx, dy;
- gradient_t *gradient = (gradient_t *)image;
- linear_gradient_t *linear = (linear_gradient_t *)image;
- uint32_t *end = buffer + width;
- pixman_gradient_walker_t walker;
-
- _pixman_gradient_walker_init (&walker, gradient, image->common.repeat);
-
- /* reference point is the center of the pixel */
- v.vector[0] = pixman_int_to_fixed (x) + pixman_fixed_1 / 2;
- v.vector[1] = pixman_int_to_fixed (y) + pixman_fixed_1 / 2;
- v.vector[2] = pixman_fixed_1;
-
- if (image->common.transform)
- {
- if (!pixman_transform_point_3d (image->common.transform, &v))
- return;
-
- unit.vector[0] = image->common.transform->matrix[0][0];
- unit.vector[1] = image->common.transform->matrix[1][0];
- unit.vector[2] = image->common.transform->matrix[2][0];
- }
- else
- {
- unit.vector[0] = pixman_fixed_1;
- unit.vector[1] = 0;
- unit.vector[2] = 0;
- }
-
- dx = linear->p2.x - linear->p1.x;
- dy = linear->p2.y - linear->p1.y;
-
- l = dx * dx + dy * dy;
-
- if (l == 0 || unit.vector[2] == 0)
- {
- /* affine transformation only */
- pixman_fixed_32_32_t t, next_inc;
- double inc;
-
- if (l == 0 || v.vector[2] == 0)
- {
- t = 0;
- inc = 0;
- }
- else
- {
- double invden, v2;
-
- invden = pixman_fixed_1 * (double) pixman_fixed_1 /
- (l * (double) v.vector[2]);
- v2 = v.vector[2] * (1. / pixman_fixed_1);
- t = ((dx * v.vector[0] + dy * v.vector[1]) -
- (dx * linear->p1.x + dy * linear->p1.y) * v2) * invden;
- inc = (dx * unit.vector[0] + dy * unit.vector[1]) * invden;
- }
- next_inc = 0;
-
- if (((pixman_fixed_32_32_t )(inc * width)) == 0)
- {
- register uint32_t color;
-
- color = _pixman_gradient_walker_pixel (&walker, t);
- while (buffer < end)
- *buffer++ = color;
- }
- else
- {
- int i;
-
- i = 0;
- while (buffer < end)
- {
- if (!mask || *mask++)
- {
- *buffer = _pixman_gradient_walker_pixel (&walker,
- t + next_inc);
- }
- i++;
- next_inc = inc * i;
- buffer++;
- }
- }
- }
- else
- {
- /* projective transformation */
- double t;
-
- t = 0;
-
- while (buffer < end)
- {
- if (!mask || *mask++)
- {
- if (v.vector[2] != 0)
- {
- double invden, v2;
-
- invden = pixman_fixed_1 * (double) pixman_fixed_1 /
- (l * (double) v.vector[2]);
- v2 = v.vector[2] * (1. / pixman_fixed_1);
- t = ((dx * v.vector[0] + dy * v.vector[1]) -
- (dx * linear->p1.x + dy * linear->p1.y) * v2) * invden;
- }
-
- *buffer = _pixman_gradient_walker_pixel (&walker, t);
- }
-
- ++buffer;
-
- v.vector[0] += unit.vector[0];
- v.vector[1] += unit.vector[1];
- v.vector[2] += unit.vector[2];
- }
- }
-}
-
-static void
-linear_gradient_property_changed (pixman_image_t *image)
-{
- image->common.get_scanline_32 = linear_gradient_get_scanline_32;
- image->common.get_scanline_64 = _pixman_image_get_scanline_generic_64;
-}
-
-PIXMAN_EXPORT pixman_image_t *
-pixman_image_create_linear_gradient (pixman_point_fixed_t * p1,
- pixman_point_fixed_t * p2,
- const pixman_gradient_stop_t *stops,
- int n_stops)
-{
- pixman_image_t *image;
- linear_gradient_t *linear;
-
- image = _pixman_image_allocate ();
-
- if (!image)
- return NULL;
-
- linear = &image->linear;
-
- if (!_pixman_init_gradient (&linear->common, stops, n_stops))
- {
- free (image);
- return NULL;
- }
-
- linear->p1 = *p1;
- linear->p2 = *p2;
-
- image->type = LINEAR;
- image->common.classify = linear_gradient_classify;
- image->common.property_changed = linear_gradient_property_changed;
-
- return image;
-}
-
+/* -*- Mode: c; c-basic-offset: 4; tab-width: 8; indent-tabs-mode: t; -*- */ +/* + * Copyright © 2000 SuSE, Inc. + * Copyright © 2007 Red Hat, Inc. + * Copyright © 2000 Keith Packard, member of The XFree86 Project, Inc. + * 2005 Lars Knoll & Zack Rusin, Trolltech + * + * Permission to use, copy, modify, distribute, and sell this software and its + * documentation for any purpose is hereby granted without fee, provided that + * the above copyright notice appear in all copies and that both that + * copyright notice and this permission notice appear in supporting + * documentation, and that the name of Keith Packard not be used in + * advertising or publicity pertaining to distribution of the software without + * specific, written prior permission. Keith Packard makes no + * representations about the suitability of this software for any purpose. It + * is provided "as is" without express or implied warranty. + * + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS + * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND + * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY + * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN + * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING + * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS + * SOFTWARE. + */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif +#include <stdlib.h> +#include "pixman-private.h" + +static pixman_bool_t +linear_gradient_is_horizontal (pixman_image_t *image, + int x, + int y, + int width, + int height) +{ + linear_gradient_t *linear = (linear_gradient_t *)image; + pixman_vector_t v; + pixman_fixed_32_32_t l; + pixman_fixed_48_16_t dx, dy; + double inc; + + if (image->common.transform) + { + /* projective transformation */ + if (image->common.transform->matrix[2][0] != 0 || + image->common.transform->matrix[2][1] != 0 || + image->common.transform->matrix[2][2] == 0) + { + return FALSE; + } + + v.vector[0] = image->common.transform->matrix[0][1]; + v.vector[1] = image->common.transform->matrix[1][1]; + v.vector[2] = image->common.transform->matrix[2][2]; + } + else + { + v.vector[0] = 0; + v.vector[1] = pixman_fixed_1; + v.vector[2] = pixman_fixed_1; + } + + dx = linear->p2.x - linear->p1.x; + dy = linear->p2.y - linear->p1.y; + + l = dx * dx + dy * dy; + + if (l == 0) + return FALSE; + + /* + * compute how much the input of the gradient walked changes + * when moving vertically through the whole image + */ + inc = height * (double) pixman_fixed_1 * pixman_fixed_1 * + (dx * v.vector[0] + dy * v.vector[1]) / + (v.vector[2] * (double) l); + + /* check that casting to integer would result in 0 */ + if (-1 < inc && inc < 1) + return TRUE; + + return FALSE; +} + +static uint32_t * +linear_get_scanline_narrow (pixman_iter_t *iter, + const uint32_t *mask) +{ + pixman_image_t *image = iter->image; + int x = iter->x; + int y = iter->y; + int width = iter->width; + uint32_t * buffer = iter->buffer; + + pixman_vector_t v, unit; + pixman_fixed_32_32_t l; + pixman_fixed_48_16_t dx, dy; + gradient_t *gradient = (gradient_t *)image; + linear_gradient_t *linear = (linear_gradient_t *)image; + uint32_t *end = buffer + width; + pixman_gradient_walker_t walker; + + _pixman_gradient_walker_init (&walker, gradient, image->common.repeat); + + /* reference point is the center of the pixel */ + v.vector[0] = pixman_int_to_fixed (x) + pixman_fixed_1 / 2; + v.vector[1] = pixman_int_to_fixed (y) + pixman_fixed_1 / 2; + v.vector[2] = pixman_fixed_1; + + if (image->common.transform) + { + if (!pixman_transform_point_3d (image->common.transform, &v)) + return iter->buffer; + + unit.vector[0] = image->common.transform->matrix[0][0]; + unit.vector[1] = image->common.transform->matrix[1][0]; + unit.vector[2] = image->common.transform->matrix[2][0]; + } + else + { + unit.vector[0] = pixman_fixed_1; + unit.vector[1] = 0; + unit.vector[2] = 0; + } + + dx = linear->p2.x - linear->p1.x; + dy = linear->p2.y - linear->p1.y; + + l = dx * dx + dy * dy; + + if (l == 0 || unit.vector[2] == 0) + { + /* affine transformation only */ + pixman_fixed_32_32_t t, next_inc; + double inc; + + if (l == 0 || v.vector[2] == 0) + { + t = 0; + inc = 0; + } + else + { + double invden, v2; + + invden = pixman_fixed_1 * (double) pixman_fixed_1 / + (l * (double) v.vector[2]); + v2 = v.vector[2] * (1. / pixman_fixed_1); + t = ((dx * v.vector[0] + dy * v.vector[1]) - + (dx * linear->p1.x + dy * linear->p1.y) * v2) * invden; + inc = (dx * unit.vector[0] + dy * unit.vector[1]) * invden; + } + next_inc = 0; + + if (((pixman_fixed_32_32_t )(inc * width)) == 0) + { + register uint32_t color; + + color = _pixman_gradient_walker_pixel (&walker, t); + while (buffer < end) + *buffer++ = color; + } + else + { + int i; + + i = 0; + while (buffer < end) + { + if (!mask || *mask++) + { + *buffer = _pixman_gradient_walker_pixel (&walker, + t + next_inc); + } + i++; + next_inc = inc * i; + buffer++; + } + } + } + else + { + /* projective transformation */ + double t; + + t = 0; + + while (buffer < end) + { + if (!mask || *mask++) + { + if (v.vector[2] != 0) + { + double invden, v2; + + invden = pixman_fixed_1 * (double) pixman_fixed_1 / + (l * (double) v.vector[2]); + v2 = v.vector[2] * (1. / pixman_fixed_1); + t = ((dx * v.vector[0] + dy * v.vector[1]) - + (dx * linear->p1.x + dy * linear->p1.y) * v2) * invden; + } + + *buffer = _pixman_gradient_walker_pixel (&walker, t); + } + + ++buffer; + + v.vector[0] += unit.vector[0]; + v.vector[1] += unit.vector[1]; + v.vector[2] += unit.vector[2]; + } + } + + iter->y++; + + return iter->buffer; +} + +static uint32_t * +linear_get_scanline_wide (pixman_iter_t *iter, const uint32_t *mask) +{ + uint32_t *buffer = linear_get_scanline_narrow (iter, NULL); + + pixman_expand ((uint64_t *)buffer, buffer, PIXMAN_a8r8g8b8, iter->width); + + return buffer; +} + +void +_pixman_linear_gradient_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, + int y, + int width, + int height, + uint8_t *buffer, + iter_flags_t flags) +{ + if (linear_gradient_is_horizontal (image, x, y, width, height)) + { + if (flags & ITER_NARROW) + linear_get_scanline_narrow (iter, NULL); + else + linear_get_scanline_wide (iter, NULL); + + iter->get_scanline = _pixman_iter_get_scanline_noop; + } + else + { + if (flags & ITER_NARROW) + iter->get_scanline = linear_get_scanline_narrow; + else + iter->get_scanline = linear_get_scanline_wide; + } +} + +PIXMAN_EXPORT pixman_image_t * +pixman_image_create_linear_gradient (pixman_point_fixed_t * p1, + pixman_point_fixed_t * p2, + const pixman_gradient_stop_t *stops, + int n_stops) +{ + pixman_image_t *image; + linear_gradient_t *linear; + + image = _pixman_image_allocate (); + + if (!image) + return NULL; + + linear = &image->linear; + + if (!_pixman_init_gradient (&linear->common, stops, n_stops)) + { + free (image); + return NULL; + } + + linear->p1 = *p1; + linear->p2 = *p2; + + image->type = LINEAR; + + return image; +} + diff --git a/pixman/pixman/pixman-mmx.c b/pixman/pixman/pixman-mmx.c index d05a185ea..6daa364fb 100644 --- a/pixman/pixman/pixman-mmx.c +++ b/pixman/pixman/pixman-mmx.c @@ -1,3378 +1,3378 @@ -/*
- * Copyright © 2004, 2005 Red Hat, Inc.
- * Copyright © 2004 Nicholas Miell
- * Copyright © 2005 Trolltech AS
- *
- * Permission to use, copy, modify, distribute, and sell this software and its
- * documentation for any purpose is hereby granted without fee, provided that
- * the above copyright notice appear in all copies and that both that
- * copyright notice and this permission notice appear in supporting
- * documentation, and that the name of Red Hat not be used in advertising or
- * publicity pertaining to distribution of the software without specific,
- * written prior permission. Red Hat makes no representations about the
- * suitability of this software for any purpose. It is provided "as is"
- * without express or implied warranty.
- *
- * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
- * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
- * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
- * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING
- * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
- * SOFTWARE.
- *
- * Author: Søren Sandmann (sandmann@redhat.com)
- * Minor Improvements: Nicholas Miell (nmiell@gmail.com)
- * MMX code paths for fbcompose.c by Lars Knoll (lars@trolltech.com)
- *
- * Based on work by Owen Taylor
- */
-
-#ifdef HAVE_CONFIG_H
-#include <config.h>
-#endif
-
-#ifdef USE_MMX
-
-#include <mmintrin.h>
-#include "pixman-private.h"
-#include "pixman-combine32.h"
-
-#define no_vERBOSE
-
-#ifdef VERBOSE
-#define CHECKPOINT() error_f ("at %s %d\n", __FUNCTION__, __LINE__)
-#else
-#define CHECKPOINT()
-#endif
-
-/* Notes about writing mmx code
- *
- * give memory operands as the second operand. If you give it as the
- * first, gcc will first load it into a register, then use that
- * register
- *
- * ie. use
- *
- * _mm_mullo_pi16 (x, mmx_constant);
- *
- * not
- *
- * _mm_mullo_pi16 (mmx_constant, x);
- *
- * Also try to minimize dependencies. i.e. when you need a value, try
- * to calculate it from a value that was calculated as early as
- * possible.
- */
-
-/* --------------- MMX primitives ------------------------------------- */
-
-#ifdef __GNUC__
-typedef uint64_t mmxdatafield;
-#else
-typedef __m64 mmxdatafield;
-/* If __m64 is defined as a struct or union, define M64_MEMBER to be the
- name of the member used to access the data */
-# ifdef _MSC_VER
-# define M64_MEMBER m64_u64
-# elif defined(__SUNPRO_C)
-# define M64_MEMBER l_
-# endif
-#endif
-
-typedef struct
-{
- mmxdatafield mmx_4x00ff;
- mmxdatafield mmx_4x0080;
- mmxdatafield mmx_565_rgb;
- mmxdatafield mmx_565_unpack_multiplier;
- mmxdatafield mmx_565_r;
- mmxdatafield mmx_565_g;
- mmxdatafield mmx_565_b;
- mmxdatafield mmx_mask_0;
- mmxdatafield mmx_mask_1;
- mmxdatafield mmx_mask_2;
- mmxdatafield mmx_mask_3;
- mmxdatafield mmx_full_alpha;
- mmxdatafield mmx_ffff0000ffff0000;
- mmxdatafield mmx_0000ffff00000000;
- mmxdatafield mmx_000000000000ffff;
-} mmx_data_t;
-
-#if defined(_MSC_VER)
-# define MMXDATA_INIT(field, val) { val ## UI64 }
-#elif defined(M64_MEMBER) /* __m64 is a struct, not an integral type */
-# define MMXDATA_INIT(field, val) field = { val ## ULL }
-#else /* __m64 is an integral type */
-# define MMXDATA_INIT(field, val) field = val ## ULL
-#endif
-
-static const mmx_data_t c =
-{
- MMXDATA_INIT (.mmx_4x00ff, 0x00ff00ff00ff00ff),
- MMXDATA_INIT (.mmx_4x0080, 0x0080008000800080),
- MMXDATA_INIT (.mmx_565_rgb, 0x000001f0003f001f),
- MMXDATA_INIT (.mmx_565_unpack_multiplier, 0x0000008404100840),
- MMXDATA_INIT (.mmx_565_r, 0x000000f800000000),
- MMXDATA_INIT (.mmx_565_g, 0x0000000000fc0000),
- MMXDATA_INIT (.mmx_565_b, 0x00000000000000f8),
- MMXDATA_INIT (.mmx_mask_0, 0xffffffffffff0000),
- MMXDATA_INIT (.mmx_mask_1, 0xffffffff0000ffff),
- MMXDATA_INIT (.mmx_mask_2, 0xffff0000ffffffff),
- MMXDATA_INIT (.mmx_mask_3, 0x0000ffffffffffff),
- MMXDATA_INIT (.mmx_full_alpha, 0x00ff000000000000),
- MMXDATA_INIT (.mmx_ffff0000ffff0000, 0xffff0000ffff0000),
- MMXDATA_INIT (.mmx_0000ffff00000000, 0x0000ffff00000000),
- MMXDATA_INIT (.mmx_000000000000ffff, 0x000000000000ffff),
-};
-
-#ifdef __GNUC__
-# ifdef __ICC
-# define MC(x) to_m64 (c.mmx_ ## x)
-# else
-# define MC(x) ((__m64)c.mmx_ ## x)
-# endif
-#else
-# define MC(x) c.mmx_ ## x
-#endif
-
-static force_inline __m64
-to_m64 (uint64_t x)
-{
-#ifdef __ICC
- return _mm_cvtsi64_m64 (x);
-#elif defined M64_MEMBER /* __m64 is a struct, not an integral type */
- __m64 res;
-
- res.M64_MEMBER = x;
- return res;
-#else /* __m64 is an integral type */
- return (__m64)x;
-#endif
-}
-
-static force_inline uint64_t
-to_uint64 (__m64 x)
-{
-#ifdef __ICC
- return _mm_cvtm64_si64 (x);
-#elif defined M64_MEMBER /* __m64 is a struct, not an integral type */
- uint64_t res = x.M64_MEMBER;
- return res;
-#else /* __m64 is an integral type */
- return (uint64_t)x;
-#endif
-}
-
-static force_inline __m64
-shift (__m64 v,
- int s)
-{
- if (s > 0)
- return _mm_slli_si64 (v, s);
- else if (s < 0)
- return _mm_srli_si64 (v, -s);
- else
- return v;
-}
-
-static force_inline __m64
-negate (__m64 mask)
-{
- return _mm_xor_si64 (mask, MC (4x00ff));
-}
-
-static force_inline __m64
-pix_multiply (__m64 a, __m64 b)
-{
- __m64 res;
-
- res = _mm_mullo_pi16 (a, b);
- res = _mm_adds_pu16 (res, MC (4x0080));
- res = _mm_adds_pu16 (res, _mm_srli_pi16 (res, 8));
- res = _mm_srli_pi16 (res, 8);
-
- return res;
-}
-
-static force_inline __m64
-pix_add (__m64 a, __m64 b)
-{
- return _mm_adds_pu8 (a, b);
-}
-
-static force_inline __m64
-expand_alpha (__m64 pixel)
-{
- __m64 t1, t2;
-
- t1 = shift (pixel, -48);
- t2 = shift (t1, 16);
- t1 = _mm_or_si64 (t1, t2);
- t2 = shift (t1, 32);
- t1 = _mm_or_si64 (t1, t2);
-
- return t1;
-}
-
-static force_inline __m64
-expand_alpha_rev (__m64 pixel)
-{
- __m64 t1, t2;
-
- /* move alpha to low 16 bits and zero the rest */
- t1 = shift (pixel, 48);
- t1 = shift (t1, -48);
-
- t2 = shift (t1, 16);
- t1 = _mm_or_si64 (t1, t2);
- t2 = shift (t1, 32);
- t1 = _mm_or_si64 (t1, t2);
-
- return t1;
-}
-
-static force_inline __m64
-invert_colors (__m64 pixel)
-{
- __m64 x, y, z;
-
- x = y = z = pixel;
-
- x = _mm_and_si64 (x, MC (ffff0000ffff0000));
- y = _mm_and_si64 (y, MC (000000000000ffff));
- z = _mm_and_si64 (z, MC (0000ffff00000000));
-
- y = shift (y, 32);
- z = shift (z, -32);
-
- x = _mm_or_si64 (x, y);
- x = _mm_or_si64 (x, z);
-
- return x;
-}
-
-static force_inline __m64
-over (__m64 src,
- __m64 srca,
- __m64 dest)
-{
- return _mm_adds_pu8 (src, pix_multiply (dest, negate (srca)));
-}
-
-static force_inline __m64
-over_rev_non_pre (__m64 src, __m64 dest)
-{
- __m64 srca = expand_alpha (src);
- __m64 srcfaaa = _mm_or_si64 (srca, MC (full_alpha));
-
- return over (pix_multiply (invert_colors (src), srcfaaa), srca, dest);
-}
-
-static force_inline __m64
-in (__m64 src, __m64 mask)
-{
- return pix_multiply (src, mask);
-}
-
-static force_inline __m64
-in_over_full_src_alpha (__m64 src, __m64 mask, __m64 dest)
-{
- src = _mm_or_si64 (src, MC (full_alpha));
-
- return over (in (src, mask), mask, dest);
-}
-
-#ifndef _MSC_VER
-static force_inline __m64
-in_over (__m64 src, __m64 srca, __m64 mask, __m64 dest)
-{
- return over (in (src, mask), pix_multiply (srca, mask), dest);
-}
-
-#else
-
-#define in_over(src, srca, mask, dest) \
- over (in (src, mask), pix_multiply (srca, mask), dest)
-
-#endif
-
-static force_inline __m64
-load8888 (uint32_t v)
-{
- return _mm_unpacklo_pi8 (_mm_cvtsi32_si64 (v), _mm_setzero_si64 ());
-}
-
-static force_inline __m64
-pack8888 (__m64 lo, __m64 hi)
-{
- return _mm_packs_pu16 (lo, hi);
-}
-
-static force_inline uint32_t
-store8888 (__m64 v)
-{
- return _mm_cvtsi64_si32 (pack8888 (v, _mm_setzero_si64 ()));
-}
-
-/* Expand 16 bits positioned at @pos (0-3) of a mmx register into
- *
- * 00RR00GG00BB
- *
- * --- Expanding 565 in the low word ---
- *
- * m = (m << (32 - 3)) | (m << (16 - 5)) | m;
- * m = m & (01f0003f001f);
- * m = m * (008404100840);
- * m = m >> 8;
- *
- * Note the trick here - the top word is shifted by another nibble to
- * avoid it bumping into the middle word
- */
-static force_inline __m64
-expand565 (__m64 pixel, int pos)
-{
- __m64 p = pixel;
- __m64 t1, t2;
-
- /* move pixel to low 16 bit and zero the rest */
- p = shift (shift (p, (3 - pos) * 16), -48);
-
- t1 = shift (p, 36 - 11);
- t2 = shift (p, 16 - 5);
-
- p = _mm_or_si64 (t1, p);
- p = _mm_or_si64 (t2, p);
- p = _mm_and_si64 (p, MC (565_rgb));
-
- pixel = _mm_mullo_pi16 (p, MC (565_unpack_multiplier));
- return _mm_srli_pi16 (pixel, 8);
-}
-
-static force_inline __m64
-expand8888 (__m64 in, int pos)
-{
- if (pos == 0)
- return _mm_unpacklo_pi8 (in, _mm_setzero_si64 ());
- else
- return _mm_unpackhi_pi8 (in, _mm_setzero_si64 ());
-}
-
-static force_inline __m64
-expandx888 (__m64 in, int pos)
-{
- return _mm_or_si64 (expand8888 (in, pos), MC (full_alpha));
-}
-
-static force_inline __m64
-pack_565 (__m64 pixel, __m64 target, int pos)
-{
- __m64 p = pixel;
- __m64 t = target;
- __m64 r, g, b;
-
- r = _mm_and_si64 (p, MC (565_r));
- g = _mm_and_si64 (p, MC (565_g));
- b = _mm_and_si64 (p, MC (565_b));
-
- r = shift (r, -(32 - 8) + pos * 16);
- g = shift (g, -(16 - 3) + pos * 16);
- b = shift (b, -(0 + 3) + pos * 16);
-
- if (pos == 0)
- t = _mm_and_si64 (t, MC (mask_0));
- else if (pos == 1)
- t = _mm_and_si64 (t, MC (mask_1));
- else if (pos == 2)
- t = _mm_and_si64 (t, MC (mask_2));
- else if (pos == 3)
- t = _mm_and_si64 (t, MC (mask_3));
-
- p = _mm_or_si64 (r, t);
- p = _mm_or_si64 (g, p);
-
- return _mm_or_si64 (b, p);
-}
-
-#ifndef _MSC_VER
-
-static force_inline __m64
-pix_add_mul (__m64 x, __m64 a, __m64 y, __m64 b)
-{
- x = pix_multiply (x, a);
- y = pix_multiply (y, b);
-
- return pix_add (x, y);
-}
-
-#else
-
-#define pix_add_mul(x, a, y, b) \
- ( x = pix_multiply (x, a), \
- y = pix_multiply (y, a), \
- pix_add (x, y) )
-
-#endif
-
-/* --------------- MMX code patch for fbcompose.c --------------------- */
-
-static force_inline uint32_t
-combine (const uint32_t *src, const uint32_t *mask)
-{
- uint32_t ssrc = *src;
-
- if (mask)
- {
- __m64 m = load8888 (*mask);
- __m64 s = load8888 (ssrc);
-
- m = expand_alpha (m);
- s = pix_multiply (s, m);
-
- ssrc = store8888 (s);
- }
-
- return ssrc;
-}
-
-static void
-mmx_combine_over_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- uint32_t ssrc = combine (src, mask);
- uint32_t a = ssrc >> 24;
-
- if (a == 0xff)
- {
- *dest = ssrc;
- }
- else if (ssrc)
- {
- __m64 s, sa;
- s = load8888 (ssrc);
- sa = expand_alpha (s);
- *dest = store8888 (over (s, sa, load8888 (*dest)));
- }
-
- ++dest;
- ++src;
- if (mask)
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_over_reverse_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- __m64 d, da;
- uint32_t s = combine (src, mask);
-
- d = load8888 (*dest);
- da = expand_alpha (d);
- *dest = store8888 (over (d, da, load8888 (s)));
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_in_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- __m64 x, a;
-
- x = load8888 (combine (src, mask));
- a = load8888 (*dest);
- a = expand_alpha (a);
- x = pix_multiply (x, a);
-
- *dest = store8888 (x);
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_in_reverse_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- __m64 x, a;
-
- x = load8888 (*dest);
- a = load8888 (combine (src, mask));
- a = expand_alpha (a);
- x = pix_multiply (x, a);
- *dest = store8888 (x);
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_out_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- __m64 x, a;
-
- x = load8888 (combine (src, mask));
- a = load8888 (*dest);
- a = expand_alpha (a);
- a = negate (a);
- x = pix_multiply (x, a);
- *dest = store8888 (x);
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_out_reverse_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- __m64 x, a;
-
- x = load8888 (*dest);
- a = load8888 (combine (src, mask));
- a = expand_alpha (a);
- a = negate (a);
- x = pix_multiply (x, a);
-
- *dest = store8888 (x);
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_atop_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- __m64 s, da, d, sia;
-
- s = load8888 (combine (src, mask));
- d = load8888 (*dest);
- sia = expand_alpha (s);
- sia = negate (sia);
- da = expand_alpha (d);
- s = pix_add_mul (s, da, d, sia);
- *dest = store8888 (s);
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_atop_reverse_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end;
-
- end = dest + width;
-
- while (dest < end)
- {
- __m64 s, dia, d, sa;
-
- s = load8888 (combine (src, mask));
- d = load8888 (*dest);
- sa = expand_alpha (s);
- dia = expand_alpha (d);
- dia = negate (dia);
- s = pix_add_mul (s, dia, d, sa);
- *dest = store8888 (s);
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_xor_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- __m64 s, dia, d, sia;
-
- s = load8888 (combine (src, mask));
- d = load8888 (*dest);
- sia = expand_alpha (s);
- dia = expand_alpha (d);
- sia = negate (sia);
- dia = negate (dia);
- s = pix_add_mul (s, dia, d, sia);
- *dest = store8888 (s);
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_add_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- __m64 s, d;
-
- s = load8888 (combine (src, mask));
- d = load8888 (*dest);
- s = pix_add (s, d);
- *dest = store8888 (s);
-
- ++dest;
- ++src;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_saturate_u (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = dest + width;
-
- while (dest < end)
- {
- uint32_t s = combine (src, mask);
- uint32_t d = *dest;
- __m64 ms = load8888 (s);
- __m64 md = load8888 (d);
- uint32_t sa = s >> 24;
- uint32_t da = ~d >> 24;
-
- if (sa > da)
- {
- __m64 msa = load8888 (DIV_UN8 (da, sa) << 24);
- msa = expand_alpha (msa);
- ms = pix_multiply (ms, msa);
- }
-
- md = pix_add (md, ms);
- *dest = store8888 (md);
-
- ++src;
- ++dest;
- if (mask)
- mask++;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_src_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
-
- s = pix_multiply (s, a);
- *dest = store8888 (s);
-
- ++src;
- ++mask;
- ++dest;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_over_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 sa = expand_alpha (s);
-
- *dest = store8888 (in_over (s, sa, a, d));
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_over_reverse_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 da = expand_alpha (d);
-
- *dest = store8888 (over (d, da, in (s, a)));
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_in_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 da = expand_alpha (d);
-
- s = pix_multiply (s, a);
- s = pix_multiply (s, da);
- *dest = store8888 (s);
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_in_reverse_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 sa = expand_alpha (s);
-
- a = pix_multiply (a, sa);
- d = pix_multiply (d, a);
- *dest = store8888 (d);
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_out_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 da = expand_alpha (d);
-
- da = negate (da);
- s = pix_multiply (s, a);
- s = pix_multiply (s, da);
- *dest = store8888 (s);
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_out_reverse_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 sa = expand_alpha (s);
-
- a = pix_multiply (a, sa);
- a = negate (a);
- d = pix_multiply (d, a);
- *dest = store8888 (d);
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_atop_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 da = expand_alpha (d);
- __m64 sa = expand_alpha (s);
-
- s = pix_multiply (s, a);
- a = pix_multiply (a, sa);
- a = negate (a);
- d = pix_add_mul (d, a, s, da);
- *dest = store8888 (d);
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_atop_reverse_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 da = expand_alpha (d);
- __m64 sa = expand_alpha (s);
-
- s = pix_multiply (s, a);
- a = pix_multiply (a, sa);
- da = negate (da);
- d = pix_add_mul (d, a, s, da);
- *dest = store8888 (d);
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_xor_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
- __m64 da = expand_alpha (d);
- __m64 sa = expand_alpha (s);
-
- s = pix_multiply (s, a);
- a = pix_multiply (a, sa);
- da = negate (da);
- a = negate (a);
- d = pix_add_mul (d, a, s, da);
- *dest = store8888 (d);
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-static void
-mmx_combine_add_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width)
-{
- const uint32_t *end = src + width;
-
- while (src < end)
- {
- __m64 a = load8888 (*mask);
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dest);
-
- s = pix_multiply (s, a);
- d = pix_add (s, d);
- *dest = store8888 (d);
-
- ++src;
- ++dest;
- ++mask;
- }
- _mm_empty ();
-}
-
-/* ------------- MMX code paths called from fbpict.c -------------------- */
-
-static void
-mmx_composite_over_n_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t src;
- uint32_t *dst_line, *dst;
- int32_t w;
- int dst_stride;
- __m64 vsrc, vsrca;
-
- CHECKPOINT ();
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- if (src == 0)
- return;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- w = width;
-
- CHECKPOINT ();
-
- while (w && (unsigned long)dst & 7)
- {
- *dst = store8888 (over (vsrc, vsrca, load8888 (*dst)));
-
- w--;
- dst++;
- }
-
- while (w >= 2)
- {
- __m64 vdest;
- __m64 dest0, dest1;
-
- vdest = *(__m64 *)dst;
-
- dest0 = over (vsrc, vsrca, expand8888 (vdest, 0));
- dest1 = over (vsrc, vsrca, expand8888 (vdest, 1));
-
- *(__m64 *)dst = pack8888 (dest0, dest1);
-
- dst += 2;
- w -= 2;
- }
-
- CHECKPOINT ();
-
- while (w)
- {
- *dst = store8888 (over (vsrc, vsrca, load8888 (*dst)));
-
- w--;
- dst++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_n_0565 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t src;
- uint16_t *dst_line, *dst;
- int32_t w;
- int dst_stride;
- __m64 vsrc, vsrca;
-
- CHECKPOINT ();
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- if (src == 0)
- return;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1);
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- w = width;
-
- CHECKPOINT ();
-
- while (w && (unsigned long)dst & 7)
- {
- uint64_t d = *dst;
- __m64 vdest = expand565 (to_m64 (d), 0);
-
- vdest = pack_565 (over (vsrc, vsrca, vdest), vdest, 0);
- *dst = to_uint64 (vdest);
-
- w--;
- dst++;
- }
-
- while (w >= 4)
- {
- __m64 vdest;
-
- vdest = *(__m64 *)dst;
-
- vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 0)), vdest, 0);
- vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 1)), vdest, 1);
- vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 2)), vdest, 2);
- vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 3)), vdest, 3);
-
- *(__m64 *)dst = vdest;
-
- dst += 4;
- w -= 4;
- }
-
- CHECKPOINT ();
-
- while (w)
- {
- uint64_t d = *dst;
- __m64 vdest = expand565 (to_m64 (d), 0);
-
- vdest = pack_565 (over (vsrc, vsrca, vdest), vdest, 0);
- *dst = to_uint64 (vdest);
-
- w--;
- dst++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_n_8888_8888_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t src, srca;
- uint32_t *dst_line;
- uint32_t *mask_line;
- int dst_stride, mask_stride;
- __m64 vsrc, vsrca;
-
- CHECKPOINT ();
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- srca = src >> 24;
- if (src == 0)
- return;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint32_t, mask_stride, mask_line, 1);
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- while (height--)
- {
- int twidth = width;
- uint32_t *p = (uint32_t *)mask_line;
- uint32_t *q = (uint32_t *)dst_line;
-
- while (twidth && (unsigned long)q & 7)
- {
- uint32_t m = *(uint32_t *)p;
-
- if (m)
- {
- __m64 vdest = load8888 (*q);
- vdest = in_over (vsrc, vsrca, load8888 (m), vdest);
- *q = store8888 (vdest);
- }
-
- twidth--;
- p++;
- q++;
- }
-
- while (twidth >= 2)
- {
- uint32_t m0, m1;
- m0 = *p;
- m1 = *(p + 1);
-
- if (m0 | m1)
- {
- __m64 dest0, dest1;
- __m64 vdest = *(__m64 *)q;
-
- dest0 = in_over (vsrc, vsrca, load8888 (m0),
- expand8888 (vdest, 0));
- dest1 = in_over (vsrc, vsrca, load8888 (m1),
- expand8888 (vdest, 1));
-
- *(__m64 *)q = pack8888 (dest0, dest1);
- }
-
- p += 2;
- q += 2;
- twidth -= 2;
- }
-
- while (twidth)
- {
- uint32_t m = *(uint32_t *)p;
-
- if (m)
- {
- __m64 vdest = load8888 (*q);
- vdest = in_over (vsrc, vsrca, load8888 (m), vdest);
- *q = store8888 (vdest);
- }
-
- twidth--;
- p++;
- q++;
- }
-
- dst_line += dst_stride;
- mask_line += mask_stride;
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_8888_n_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t *dst_line, *dst;
- uint32_t *src_line, *src;
- uint32_t mask;
- __m64 vmask;
- int dst_stride, src_stride;
- int32_t w;
- __m64 srca;
-
- CHECKPOINT ();
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
-
- mask = _pixman_image_get_solid (mask_image, dst_image->bits.format);
- mask &= 0xff000000;
- mask = mask | mask >> 8 | mask >> 16 | mask >> 24;
- vmask = load8888 (mask);
- srca = MC (4x00ff);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- while (w && (unsigned long)dst & 7)
- {
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dst);
-
- *dst = store8888 (in_over (s, expand_alpha (s), vmask, d));
-
- w--;
- dst++;
- src++;
- }
-
- while (w >= 2)
- {
- __m64 vs = *(__m64 *)src;
- __m64 vd = *(__m64 *)dst;
- __m64 vsrc0 = expand8888 (vs, 0);
- __m64 vsrc1 = expand8888 (vs, 1);
-
- *(__m64 *)dst = pack8888 (
- in_over (vsrc0, expand_alpha (vsrc0), vmask, expand8888 (vd, 0)),
- in_over (vsrc1, expand_alpha (vsrc1), vmask, expand8888 (vd, 1)));
-
- w -= 2;
- dst += 2;
- src += 2;
- }
-
- while (w)
- {
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dst);
-
- *dst = store8888 (in_over (s, expand_alpha (s), vmask, d));
-
- w--;
- dst++;
- src++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_x888_n_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t *dst_line, *dst;
- uint32_t *src_line, *src;
- uint32_t mask;
- __m64 vmask;
- int dst_stride, src_stride;
- int32_t w;
- __m64 srca;
-
- CHECKPOINT ();
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
- mask = _pixman_image_get_solid (mask_image, dst_image->bits.format);
-
- mask &= 0xff000000;
- mask = mask | mask >> 8 | mask >> 16 | mask >> 24;
- vmask = load8888 (mask);
- srca = MC (4x00ff);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- while (w && (unsigned long)dst & 7)
- {
- __m64 s = load8888 (*src | 0xff000000);
- __m64 d = load8888 (*dst);
-
- *dst = store8888 (in_over (s, srca, vmask, d));
-
- w--;
- dst++;
- src++;
- }
-
- while (w >= 16)
- {
- __m64 vd0 = *(__m64 *)(dst + 0);
- __m64 vd1 = *(__m64 *)(dst + 2);
- __m64 vd2 = *(__m64 *)(dst + 4);
- __m64 vd3 = *(__m64 *)(dst + 6);
- __m64 vd4 = *(__m64 *)(dst + 8);
- __m64 vd5 = *(__m64 *)(dst + 10);
- __m64 vd6 = *(__m64 *)(dst + 12);
- __m64 vd7 = *(__m64 *)(dst + 14);
-
- __m64 vs0 = *(__m64 *)(src + 0);
- __m64 vs1 = *(__m64 *)(src + 2);
- __m64 vs2 = *(__m64 *)(src + 4);
- __m64 vs3 = *(__m64 *)(src + 6);
- __m64 vs4 = *(__m64 *)(src + 8);
- __m64 vs5 = *(__m64 *)(src + 10);
- __m64 vs6 = *(__m64 *)(src + 12);
- __m64 vs7 = *(__m64 *)(src + 14);
-
- vd0 = pack8888 (
- in_over (expandx888 (vs0, 0), srca, vmask, expand8888 (vd0, 0)),
- in_over (expandx888 (vs0, 1), srca, vmask, expand8888 (vd0, 1)));
-
- vd1 = pack8888 (
- in_over (expandx888 (vs1, 0), srca, vmask, expand8888 (vd1, 0)),
- in_over (expandx888 (vs1, 1), srca, vmask, expand8888 (vd1, 1)));
-
- vd2 = pack8888 (
- in_over (expandx888 (vs2, 0), srca, vmask, expand8888 (vd2, 0)),
- in_over (expandx888 (vs2, 1), srca, vmask, expand8888 (vd2, 1)));
-
- vd3 = pack8888 (
- in_over (expandx888 (vs3, 0), srca, vmask, expand8888 (vd3, 0)),
- in_over (expandx888 (vs3, 1), srca, vmask, expand8888 (vd3, 1)));
-
- vd4 = pack8888 (
- in_over (expandx888 (vs4, 0), srca, vmask, expand8888 (vd4, 0)),
- in_over (expandx888 (vs4, 1), srca, vmask, expand8888 (vd4, 1)));
-
- vd5 = pack8888 (
- in_over (expandx888 (vs5, 0), srca, vmask, expand8888 (vd5, 0)),
- in_over (expandx888 (vs5, 1), srca, vmask, expand8888 (vd5, 1)));
-
- vd6 = pack8888 (
- in_over (expandx888 (vs6, 0), srca, vmask, expand8888 (vd6, 0)),
- in_over (expandx888 (vs6, 1), srca, vmask, expand8888 (vd6, 1)));
-
- vd7 = pack8888 (
- in_over (expandx888 (vs7, 0), srca, vmask, expand8888 (vd7, 0)),
- in_over (expandx888 (vs7, 1), srca, vmask, expand8888 (vd7, 1)));
-
- *(__m64 *)(dst + 0) = vd0;
- *(__m64 *)(dst + 2) = vd1;
- *(__m64 *)(dst + 4) = vd2;
- *(__m64 *)(dst + 6) = vd3;
- *(__m64 *)(dst + 8) = vd4;
- *(__m64 *)(dst + 10) = vd5;
- *(__m64 *)(dst + 12) = vd6;
- *(__m64 *)(dst + 14) = vd7;
-
- w -= 16;
- dst += 16;
- src += 16;
- }
-
- while (w)
- {
- __m64 s = load8888 (*src | 0xff000000);
- __m64 d = load8888 (*dst);
-
- *dst = store8888 (in_over (s, srca, vmask, d));
-
- w--;
- dst++;
- src++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_8888_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t *dst_line, *dst;
- uint32_t *src_line, *src;
- uint32_t s;
- int dst_stride, src_stride;
- uint8_t a;
- int32_t w;
-
- CHECKPOINT ();
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- while (w--)
- {
- s = *src++;
- a = s >> 24;
-
- if (a == 0xff)
- {
- *dst = s;
- }
- else if (s)
- {
- __m64 ms, sa;
- ms = load8888 (s);
- sa = expand_alpha (ms);
- *dst = store8888 (over (ms, sa, load8888 (*dst)));
- }
-
- dst++;
- }
- }
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_8888_0565 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint16_t *dst_line, *dst;
- uint32_t *src_line, *src;
- int dst_stride, src_stride;
- int32_t w;
-
- CHECKPOINT ();
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
-
-#if 0
- /* FIXME */
- assert (src_image->drawable == mask_image->drawable);
-#endif
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- CHECKPOINT ();
-
- while (w && (unsigned long)dst & 7)
- {
- __m64 vsrc = load8888 (*src);
- uint64_t d = *dst;
- __m64 vdest = expand565 (to_m64 (d), 0);
-
- vdest = pack_565 (
- over (vsrc, expand_alpha (vsrc), vdest), vdest, 0);
-
- *dst = to_uint64 (vdest);
-
- w--;
- dst++;
- src++;
- }
-
- CHECKPOINT ();
-
- while (w >= 4)
- {
- __m64 vsrc0, vsrc1, vsrc2, vsrc3;
- __m64 vdest;
-
- vsrc0 = load8888 (*(src + 0));
- vsrc1 = load8888 (*(src + 1));
- vsrc2 = load8888 (*(src + 2));
- vsrc3 = load8888 (*(src + 3));
-
- vdest = *(__m64 *)dst;
-
- vdest = pack_565 (over (vsrc0, expand_alpha (vsrc0), expand565 (vdest, 0)), vdest, 0);
- vdest = pack_565 (over (vsrc1, expand_alpha (vsrc1), expand565 (vdest, 1)), vdest, 1);
- vdest = pack_565 (over (vsrc2, expand_alpha (vsrc2), expand565 (vdest, 2)), vdest, 2);
- vdest = pack_565 (over (vsrc3, expand_alpha (vsrc3), expand565 (vdest, 3)), vdest, 3);
-
- *(__m64 *)dst = vdest;
-
- w -= 4;
- dst += 4;
- src += 4;
- }
-
- CHECKPOINT ();
-
- while (w)
- {
- __m64 vsrc = load8888 (*src);
- uint64_t d = *dst;
- __m64 vdest = expand565 (to_m64 (d), 0);
-
- vdest = pack_565 (over (vsrc, expand_alpha (vsrc), vdest), vdest, 0);
-
- *dst = to_uint64 (vdest);
-
- w--;
- dst++;
- src++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_n_8_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t src, srca;
- uint32_t *dst_line, *dst;
- uint8_t *mask_line, *mask;
- int dst_stride, mask_stride;
- int32_t w;
- __m64 vsrc, vsrca;
- uint64_t srcsrc;
-
- CHECKPOINT ();
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- srca = src >> 24;
- if (src == 0)
- return;
-
- srcsrc = (uint64_t)src << 32 | src;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1);
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- mask = mask_line;
- mask_line += mask_stride;
- w = width;
-
- CHECKPOINT ();
-
- while (w && (unsigned long)dst & 7)
- {
- uint64_t m = *mask;
-
- if (m)
- {
- __m64 vdest = in_over (vsrc, vsrca,
- expand_alpha_rev (to_m64 (m)),
- load8888 (*dst));
-
- *dst = store8888 (vdest);
- }
-
- w--;
- mask++;
- dst++;
- }
-
- CHECKPOINT ();
-
- while (w >= 2)
- {
- uint64_t m0, m1;
-
- m0 = *mask;
- m1 = *(mask + 1);
-
- if (srca == 0xff && (m0 & m1) == 0xff)
- {
- *(uint64_t *)dst = srcsrc;
- }
- else if (m0 | m1)
- {
- __m64 vdest;
- __m64 dest0, dest1;
-
- vdest = *(__m64 *)dst;
-
- dest0 = in_over (vsrc, vsrca, expand_alpha_rev (to_m64 (m0)),
- expand8888 (vdest, 0));
- dest1 = in_over (vsrc, vsrca, expand_alpha_rev (to_m64 (m1)),
- expand8888 (vdest, 1));
-
- *(__m64 *)dst = pack8888 (dest0, dest1);
- }
-
- mask += 2;
- dst += 2;
- w -= 2;
- }
-
- CHECKPOINT ();
-
- while (w)
- {
- uint64_t m = *mask;
-
- if (m)
- {
- __m64 vdest = load8888 (*dst);
-
- vdest = in_over (
- vsrc, vsrca, expand_alpha_rev (to_m64 (m)), vdest);
- *dst = store8888 (vdest);
- }
-
- w--;
- mask++;
- dst++;
- }
- }
-
- _mm_empty ();
-}
-
-pixman_bool_t
-pixman_fill_mmx (uint32_t *bits,
- int stride,
- int bpp,
- int x,
- int y,
- int width,
- int height,
- uint32_t xor)
-{
- uint64_t fill;
- __m64 vfill;
- uint32_t byte_width;
- uint8_t *byte_line;
-
-#ifdef __GNUC__
- __m64 v1, v2, v3, v4, v5, v6, v7;
-#endif
-
- if (bpp != 16 && bpp != 32 && bpp != 8)
- return FALSE;
-
- if (bpp == 8)
- {
- stride = stride * (int) sizeof (uint32_t) / 1;
- byte_line = (uint8_t *)(((uint8_t *)bits) + stride * y + x);
- byte_width = width;
- stride *= 1;
- xor = (xor & 0xff) * 0x01010101;
- }
- else if (bpp == 16)
- {
- stride = stride * (int) sizeof (uint32_t) / 2;
- byte_line = (uint8_t *)(((uint16_t *)bits) + stride * y + x);
- byte_width = 2 * width;
- stride *= 2;
- xor = (xor & 0xffff) * 0x00010001;
- }
- else
- {
- stride = stride * (int) sizeof (uint32_t) / 4;
- byte_line = (uint8_t *)(((uint32_t *)bits) + stride * y + x);
- byte_width = 4 * width;
- stride *= 4;
- }
-
- fill = ((uint64_t)xor << 32) | xor;
- vfill = to_m64 (fill);
-
-#ifdef __GNUC__
- __asm__ (
- "movq %7, %0\n"
- "movq %7, %1\n"
- "movq %7, %2\n"
- "movq %7, %3\n"
- "movq %7, %4\n"
- "movq %7, %5\n"
- "movq %7, %6\n"
- : "=&y" (v1), "=&y" (v2), "=&y" (v3),
- "=&y" (v4), "=&y" (v5), "=&y" (v6), "=y" (v7)
- : "y" (vfill));
-#endif
-
- while (height--)
- {
- int w;
- uint8_t *d = byte_line;
-
- byte_line += stride;
- w = byte_width;
-
- while (w >= 1 && ((unsigned long)d & 1))
- {
- *(uint8_t *)d = (xor & 0xff);
- w--;
- d++;
- }
-
- while (w >= 2 && ((unsigned long)d & 3))
- {
- *(uint16_t *)d = xor;
- w -= 2;
- d += 2;
- }
-
- while (w >= 4 && ((unsigned long)d & 7))
- {
- *(uint32_t *)d = xor;
-
- w -= 4;
- d += 4;
- }
-
- while (w >= 64)
- {
-#ifdef __GNUC__
- __asm__ (
- "movq %1, (%0)\n"
- "movq %2, 8(%0)\n"
- "movq %3, 16(%0)\n"
- "movq %4, 24(%0)\n"
- "movq %5, 32(%0)\n"
- "movq %6, 40(%0)\n"
- "movq %7, 48(%0)\n"
- "movq %8, 56(%0)\n"
- :
- : "r" (d),
- "y" (vfill), "y" (v1), "y" (v2), "y" (v3),
- "y" (v4), "y" (v5), "y" (v6), "y" (v7)
- : "memory");
-#else
- *(__m64*) (d + 0) = vfill;
- *(__m64*) (d + 8) = vfill;
- *(__m64*) (d + 16) = vfill;
- *(__m64*) (d + 24) = vfill;
- *(__m64*) (d + 32) = vfill;
- *(__m64*) (d + 40) = vfill;
- *(__m64*) (d + 48) = vfill;
- *(__m64*) (d + 56) = vfill;
-#endif
- w -= 64;
- d += 64;
- }
-
- while (w >= 4)
- {
- *(uint32_t *)d = xor;
-
- w -= 4;
- d += 4;
- }
- while (w >= 2)
- {
- *(uint16_t *)d = xor;
- w -= 2;
- d += 2;
- }
- while (w >= 1)
- {
- *(uint8_t *)d = (xor & 0xff);
- w--;
- d++;
- }
-
- }
-
- _mm_empty ();
- return TRUE;
-}
-
-static void
-mmx_composite_src_n_8_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t src, srca;
- uint32_t *dst_line, *dst;
- uint8_t *mask_line, *mask;
- int dst_stride, mask_stride;
- int32_t w;
- __m64 vsrc, vsrca;
- uint64_t srcsrc;
-
- CHECKPOINT ();
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- srca = src >> 24;
- if (src == 0)
- {
- pixman_fill_mmx (dst_image->bits.bits, dst_image->bits.rowstride,
- PIXMAN_FORMAT_BPP (dst_image->bits.format),
- dest_x, dest_y, width, height, 0);
- return;
- }
-
- srcsrc = (uint64_t)src << 32 | src;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1);
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- mask = mask_line;
- mask_line += mask_stride;
- w = width;
-
- CHECKPOINT ();
-
- while (w && (unsigned long)dst & 7)
- {
- uint64_t m = *mask;
-
- if (m)
- {
- __m64 vdest = in (vsrc, expand_alpha_rev (to_m64 (m)));
-
- *dst = store8888 (vdest);
- }
- else
- {
- *dst = 0;
- }
-
- w--;
- mask++;
- dst++;
- }
-
- CHECKPOINT ();
-
- while (w >= 2)
- {
- uint64_t m0, m1;
- m0 = *mask;
- m1 = *(mask + 1);
-
- if (srca == 0xff && (m0 & m1) == 0xff)
- {
- *(uint64_t *)dst = srcsrc;
- }
- else if (m0 | m1)
- {
- __m64 vdest;
- __m64 dest0, dest1;
-
- vdest = *(__m64 *)dst;
-
- dest0 = in (vsrc, expand_alpha_rev (to_m64 (m0)));
- dest1 = in (vsrc, expand_alpha_rev (to_m64 (m1)));
-
- *(__m64 *)dst = pack8888 (dest0, dest1);
- }
- else
- {
- *(uint64_t *)dst = 0;
- }
-
- mask += 2;
- dst += 2;
- w -= 2;
- }
-
- CHECKPOINT ();
-
- while (w)
- {
- uint64_t m = *mask;
-
- if (m)
- {
- __m64 vdest = load8888 (*dst);
-
- vdest = in (vsrc, expand_alpha_rev (to_m64 (m)));
- *dst = store8888 (vdest);
- }
- else
- {
- *dst = 0;
- }
-
- w--;
- mask++;
- dst++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_n_8_0565 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t src, srca;
- uint16_t *dst_line, *dst;
- uint8_t *mask_line, *mask;
- int dst_stride, mask_stride;
- int32_t w;
- __m64 vsrc, vsrca, tmp;
- uint64_t srcsrcsrcsrc, src16;
-
- CHECKPOINT ();
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- srca = src >> 24;
- if (src == 0)
- return;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1);
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- tmp = pack_565 (vsrc, _mm_setzero_si64 (), 0);
- src16 = to_uint64 (tmp);
-
- srcsrcsrcsrc =
- (uint64_t)src16 << 48 | (uint64_t)src16 << 32 |
- (uint64_t)src16 << 16 | (uint64_t)src16;
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- mask = mask_line;
- mask_line += mask_stride;
- w = width;
-
- CHECKPOINT ();
-
- while (w && (unsigned long)dst & 7)
- {
- uint64_t m = *mask;
-
- if (m)
- {
- uint64_t d = *dst;
- __m64 vd = to_m64 (d);
- __m64 vdest = in_over (
- vsrc, vsrca, expand_alpha_rev (to_m64 (m)), expand565 (vd, 0));
-
- vd = pack_565 (vdest, _mm_setzero_si64 (), 0);
- *dst = to_uint64 (vd);
- }
-
- w--;
- mask++;
- dst++;
- }
-
- CHECKPOINT ();
-
- while (w >= 4)
- {
- uint64_t m0, m1, m2, m3;
- m0 = *mask;
- m1 = *(mask + 1);
- m2 = *(mask + 2);
- m3 = *(mask + 3);
-
- if (srca == 0xff && (m0 & m1 & m2 & m3) == 0xff)
- {
- *(uint64_t *)dst = srcsrcsrcsrc;
- }
- else if (m0 | m1 | m2 | m3)
- {
- __m64 vdest;
- __m64 vm0, vm1, vm2, vm3;
-
- vdest = *(__m64 *)dst;
-
- vm0 = to_m64 (m0);
- vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm0),
- expand565 (vdest, 0)), vdest, 0);
- vm1 = to_m64 (m1);
- vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm1),
- expand565 (vdest, 1)), vdest, 1);
- vm2 = to_m64 (m2);
- vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm2),
- expand565 (vdest, 2)), vdest, 2);
- vm3 = to_m64 (m3);
- vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm3),
- expand565 (vdest, 3)), vdest, 3);
-
- *(__m64 *)dst = vdest;
- }
-
- w -= 4;
- mask += 4;
- dst += 4;
- }
-
- CHECKPOINT ();
-
- while (w)
- {
- uint64_t m = *mask;
-
- if (m)
- {
- uint64_t d = *dst;
- __m64 vd = to_m64 (d);
- __m64 vdest = in_over (vsrc, vsrca, expand_alpha_rev (to_m64 (m)),
- expand565 (vd, 0));
- vd = pack_565 (vdest, _mm_setzero_si64 (), 0);
- *dst = to_uint64 (vd);
- }
-
- w--;
- mask++;
- dst++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_pixbuf_0565 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint16_t *dst_line, *dst;
- uint32_t *src_line, *src;
- int dst_stride, src_stride;
- int32_t w;
-
- CHECKPOINT ();
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
-
-#if 0
- /* FIXME */
- assert (src_image->drawable == mask_image->drawable);
-#endif
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- CHECKPOINT ();
-
- while (w && (unsigned long)dst & 7)
- {
- __m64 vsrc = load8888 (*src);
- uint64_t d = *dst;
- __m64 vdest = expand565 (to_m64 (d), 0);
-
- vdest = pack_565 (over_rev_non_pre (vsrc, vdest), vdest, 0);
-
- *dst = to_uint64 (vdest);
-
- w--;
- dst++;
- src++;
- }
-
- CHECKPOINT ();
-
- while (w >= 4)
- {
- uint32_t s0, s1, s2, s3;
- unsigned char a0, a1, a2, a3;
-
- s0 = *src;
- s1 = *(src + 1);
- s2 = *(src + 2);
- s3 = *(src + 3);
-
- a0 = (s0 >> 24);
- a1 = (s1 >> 24);
- a2 = (s2 >> 24);
- a3 = (s3 >> 24);
-
- if ((a0 & a1 & a2 & a3) == 0xFF)
- {
- __m64 vdest;
- vdest = pack_565 (invert_colors (load8888 (s0)), _mm_setzero_si64 (), 0);
- vdest = pack_565 (invert_colors (load8888 (s1)), vdest, 1);
- vdest = pack_565 (invert_colors (load8888 (s2)), vdest, 2);
- vdest = pack_565 (invert_colors (load8888 (s3)), vdest, 3);
-
- *(__m64 *)dst = vdest;
- }
- else if (s0 | s1 | s2 | s3)
- {
- __m64 vdest = *(__m64 *)dst;
-
- vdest = pack_565 (over_rev_non_pre (load8888 (s0), expand565 (vdest, 0)), vdest, 0);
- vdest = pack_565 (over_rev_non_pre (load8888 (s1), expand565 (vdest, 1)), vdest, 1);
- vdest = pack_565 (over_rev_non_pre (load8888 (s2), expand565 (vdest, 2)), vdest, 2);
- vdest = pack_565 (over_rev_non_pre (load8888 (s3), expand565 (vdest, 3)), vdest, 3);
-
- *(__m64 *)dst = vdest;
- }
-
- w -= 4;
- dst += 4;
- src += 4;
- }
-
- CHECKPOINT ();
-
- while (w)
- {
- __m64 vsrc = load8888 (*src);
- uint64_t d = *dst;
- __m64 vdest = expand565 (to_m64 (d), 0);
-
- vdest = pack_565 (over_rev_non_pre (vsrc, vdest), vdest, 0);
-
- *dst = to_uint64 (vdest);
-
- w--;
- dst++;
- src++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_pixbuf_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t *dst_line, *dst;
- uint32_t *src_line, *src;
- int dst_stride, src_stride;
- int32_t w;
-
- CHECKPOINT ();
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
-
-#if 0
- /* FIXME */
- assert (src_image->drawable == mask_image->drawable);
-#endif
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- while (w && (unsigned long)dst & 7)
- {
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dst);
-
- *dst = store8888 (over_rev_non_pre (s, d));
-
- w--;
- dst++;
- src++;
- }
-
- while (w >= 2)
- {
- uint64_t s0, s1;
- unsigned char a0, a1;
- __m64 d0, d1;
-
- s0 = *src;
- s1 = *(src + 1);
-
- a0 = (s0 >> 24);
- a1 = (s1 >> 24);
-
- if ((a0 & a1) == 0xFF)
- {
- d0 = invert_colors (load8888 (s0));
- d1 = invert_colors (load8888 (s1));
-
- *(__m64 *)dst = pack8888 (d0, d1);
- }
- else if (s0 | s1)
- {
- __m64 vdest = *(__m64 *)dst;
-
- d0 = over_rev_non_pre (load8888 (s0), expand8888 (vdest, 0));
- d1 = over_rev_non_pre (load8888 (s1), expand8888 (vdest, 1));
-
- *(__m64 *)dst = pack8888 (d0, d1);
- }
-
- w -= 2;
- dst += 2;
- src += 2;
- }
-
- while (w)
- {
- __m64 s = load8888 (*src);
- __m64 d = load8888 (*dst);
-
- *dst = store8888 (over_rev_non_pre (s, d));
-
- w--;
- dst++;
- src++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_over_n_8888_0565_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t src, srca;
- uint16_t *dst_line;
- uint32_t *mask_line;
- int dst_stride, mask_stride;
- __m64 vsrc, vsrca;
-
- CHECKPOINT ();
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- srca = src >> 24;
- if (src == 0)
- return;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint32_t, mask_stride, mask_line, 1);
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- while (height--)
- {
- int twidth = width;
- uint32_t *p = (uint32_t *)mask_line;
- uint16_t *q = (uint16_t *)dst_line;
-
- while (twidth && ((unsigned long)q & 7))
- {
- uint32_t m = *(uint32_t *)p;
-
- if (m)
- {
- uint64_t d = *q;
- __m64 vdest = expand565 (to_m64 (d), 0);
- vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m), vdest), vdest, 0);
- *q = to_uint64 (vdest);
- }
-
- twidth--;
- p++;
- q++;
- }
-
- while (twidth >= 4)
- {
- uint32_t m0, m1, m2, m3;
-
- m0 = *p;
- m1 = *(p + 1);
- m2 = *(p + 2);
- m3 = *(p + 3);
-
- if ((m0 | m1 | m2 | m3))
- {
- __m64 vdest = *(__m64 *)q;
-
- vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m0), expand565 (vdest, 0)), vdest, 0);
- vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m1), expand565 (vdest, 1)), vdest, 1);
- vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m2), expand565 (vdest, 2)), vdest, 2);
- vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m3), expand565 (vdest, 3)), vdest, 3);
-
- *(__m64 *)q = vdest;
- }
- twidth -= 4;
- p += 4;
- q += 4;
- }
-
- while (twidth)
- {
- uint32_t m;
-
- m = *(uint32_t *)p;
- if (m)
- {
- uint64_t d = *q;
- __m64 vdest = expand565 (to_m64 (d), 0);
- vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m), vdest), vdest, 0);
- *q = to_uint64 (vdest);
- }
-
- twidth--;
- p++;
- q++;
- }
-
- mask_line += mask_stride;
- dst_line += dst_stride;
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_in_n_8_8 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint8_t *dst_line, *dst;
- uint8_t *mask_line, *mask;
- int dst_stride, mask_stride;
- int32_t w;
- uint32_t src;
- uint8_t sa;
- __m64 vsrc, vsrca;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1);
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- sa = src >> 24;
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- mask = mask_line;
- mask_line += mask_stride;
- w = width;
-
- if ((((unsigned long)dst_image & 3) == 0) &&
- (((unsigned long)src_image & 3) == 0))
- {
- while (w >= 4)
- {
- uint32_t m;
- __m64 vmask;
- __m64 vdest;
-
- m = 0;
-
- vmask = load8888 (*(uint32_t *)mask);
- vdest = load8888 (*(uint32_t *)dst);
-
- *(uint32_t *)dst = store8888 (in (in (vsrca, vmask), vdest));
-
- dst += 4;
- mask += 4;
- w -= 4;
- }
- }
-
- while (w--)
- {
- uint16_t tmp;
- uint8_t a;
- uint32_t m, d;
-
- a = *mask++;
- d = *dst;
-
- m = MUL_UN8 (sa, a, tmp);
- d = MUL_UN8 (m, d, tmp);
-
- *dst++ = d;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_in_8_8 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint8_t *dst_line, *dst;
- uint8_t *src_line, *src;
- int src_stride, dst_stride;
- int32_t w;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint8_t, src_stride, src_line, 1);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- if ((((unsigned long)dst_image & 3) == 0) &&
- (((unsigned long)src_image & 3) == 0))
- {
- while (w >= 4)
- {
- uint32_t *s = (uint32_t *)src;
- uint32_t *d = (uint32_t *)dst;
-
- *d = store8888 (in (load8888 (*s), load8888 (*d)));
-
- w -= 4;
- dst += 4;
- src += 4;
- }
- }
-
- while (w--)
- {
- uint8_t s, d;
- uint16_t tmp;
-
- s = *src;
- d = *dst;
-
- *dst = MUL_UN8 (s, d, tmp);
-
- src++;
- dst++;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_add_n_8_8 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint8_t *dst_line, *dst;
- uint8_t *mask_line, *mask;
- int dst_stride, mask_stride;
- int32_t w;
- uint32_t src;
- uint8_t sa;
- __m64 vsrc, vsrca;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1);
-
- src = _pixman_image_get_solid (src_image, dst_image->bits.format);
-
- sa = src >> 24;
-
- if (src == 0)
- return;
-
- vsrc = load8888 (src);
- vsrca = expand_alpha (vsrc);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- mask = mask_line;
- mask_line += mask_stride;
- w = width;
-
- if ((((unsigned long)mask_image & 3) == 0) &&
- (((unsigned long)dst_image & 3) == 0))
- {
- while (w >= 4)
- {
- __m64 vmask = load8888 (*(uint32_t *)mask);
- __m64 vdest = load8888 (*(uint32_t *)dst);
-
- *(uint32_t *)dst = store8888 (_mm_adds_pu8 (in (vsrca, vmask), vdest));
-
- w -= 4;
- dst += 4;
- mask += 4;
- }
- }
-
- while (w--)
- {
- uint16_t tmp;
- uint16_t a;
- uint32_t m, d;
- uint32_t r;
-
- a = *mask++;
- d = *dst;
-
- m = MUL_UN8 (sa, a, tmp);
- r = ADD_UN8 (m, d, tmp);
-
- *dst++ = r;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_add_8_8 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint8_t *dst_line, *dst;
- uint8_t *src_line, *src;
- int dst_stride, src_stride;
- int32_t w;
- uint8_t s, d;
- uint16_t t;
-
- CHECKPOINT ();
-
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint8_t, src_stride, src_line, 1);
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- while (w && (unsigned long)dst & 7)
- {
- s = *src;
- d = *dst;
- t = d + s;
- s = t | (0 - (t >> 8));
- *dst = s;
-
- dst++;
- src++;
- w--;
- }
-
- while (w >= 8)
- {
- *(__m64*)dst = _mm_adds_pu8 (*(__m64*)src, *(__m64*)dst);
- dst += 8;
- src += 8;
- w -= 8;
- }
-
- while (w)
- {
- s = *src;
- d = *dst;
- t = d + s;
- s = t | (0 - (t >> 8));
- *dst = s;
-
- dst++;
- src++;
- w--;
- }
- }
-
- _mm_empty ();
-}
-
-static void
-mmx_composite_add_8888_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- __m64 dst64;
- uint32_t *dst_line, *dst;
- uint32_t *src_line, *src;
- int dst_stride, src_stride;
- int32_t w;
-
- CHECKPOINT ();
-
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
-
- while (height--)
- {
- dst = dst_line;
- dst_line += dst_stride;
- src = src_line;
- src_line += src_stride;
- w = width;
-
- while (w && (unsigned long)dst & 7)
- {
- *dst = _mm_cvtsi64_si32 (_mm_adds_pu8 (_mm_cvtsi32_si64 (*src),
- _mm_cvtsi32_si64 (*dst)));
- dst++;
- src++;
- w--;
- }
-
- while (w >= 2)
- {
- dst64 = _mm_adds_pu8 (*(__m64*)src, *(__m64*)dst);
- *(uint64_t*)dst = to_uint64 (dst64);
- dst += 2;
- src += 2;
- w -= 2;
- }
-
- if (w)
- {
- *dst = _mm_cvtsi64_si32 (_mm_adds_pu8 (_mm_cvtsi32_si64 (*src),
- _mm_cvtsi32_si64 (*dst)));
-
- }
- }
-
- _mm_empty ();
-}
-
-static pixman_bool_t
-pixman_blt_mmx (uint32_t *src_bits,
- uint32_t *dst_bits,
- int src_stride,
- int dst_stride,
- int src_bpp,
- int dst_bpp,
- int src_x,
- int src_y,
- int dst_x,
- int dst_y,
- int width,
- int height)
-{
- uint8_t * src_bytes;
- uint8_t * dst_bytes;
- int byte_width;
-
- if (src_bpp != dst_bpp)
- return FALSE;
-
- if (src_bpp == 16)
- {
- src_stride = src_stride * (int) sizeof (uint32_t) / 2;
- dst_stride = dst_stride * (int) sizeof (uint32_t) / 2;
- src_bytes = (uint8_t *)(((uint16_t *)src_bits) + src_stride * (src_y) + (src_x));
- dst_bytes = (uint8_t *)(((uint16_t *)dst_bits) + dst_stride * (dst_y) + (dst_x));
- byte_width = 2 * width;
- src_stride *= 2;
- dst_stride *= 2;
- }
- else if (src_bpp == 32)
- {
- src_stride = src_stride * (int) sizeof (uint32_t) / 4;
- dst_stride = dst_stride * (int) sizeof (uint32_t) / 4;
- src_bytes = (uint8_t *)(((uint32_t *)src_bits) + src_stride * (src_y) + (src_x));
- dst_bytes = (uint8_t *)(((uint32_t *)dst_bits) + dst_stride * (dst_y) + (dst_x));
- byte_width = 4 * width;
- src_stride *= 4;
- dst_stride *= 4;
- }
- else
- {
- return FALSE;
- }
-
- while (height--)
- {
- int w;
- uint8_t *s = src_bytes;
- uint8_t *d = dst_bytes;
- src_bytes += src_stride;
- dst_bytes += dst_stride;
- w = byte_width;
-
- while (w >= 2 && ((unsigned long)d & 3))
- {
- *(uint16_t *)d = *(uint16_t *)s;
- w -= 2;
- s += 2;
- d += 2;
- }
-
- while (w >= 4 && ((unsigned long)d & 7))
- {
- *(uint32_t *)d = *(uint32_t *)s;
-
- w -= 4;
- s += 4;
- d += 4;
- }
-
- while (w >= 64)
- {
-#if defined (__GNUC__) || (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x590))
- __asm__ (
- "movq (%1), %%mm0\n"
- "movq 8(%1), %%mm1\n"
- "movq 16(%1), %%mm2\n"
- "movq 24(%1), %%mm3\n"
- "movq 32(%1), %%mm4\n"
- "movq 40(%1), %%mm5\n"
- "movq 48(%1), %%mm6\n"
- "movq 56(%1), %%mm7\n"
-
- "movq %%mm0, (%0)\n"
- "movq %%mm1, 8(%0)\n"
- "movq %%mm2, 16(%0)\n"
- "movq %%mm3, 24(%0)\n"
- "movq %%mm4, 32(%0)\n"
- "movq %%mm5, 40(%0)\n"
- "movq %%mm6, 48(%0)\n"
- "movq %%mm7, 56(%0)\n"
- :
- : "r" (d), "r" (s)
- : "memory",
- "%mm0", "%mm1", "%mm2", "%mm3",
- "%mm4", "%mm5", "%mm6", "%mm7");
-#else
- __m64 v0 = *(__m64 *)(s + 0);
- __m64 v1 = *(__m64 *)(s + 8);
- __m64 v2 = *(__m64 *)(s + 16);
- __m64 v3 = *(__m64 *)(s + 24);
- __m64 v4 = *(__m64 *)(s + 32);
- __m64 v5 = *(__m64 *)(s + 40);
- __m64 v6 = *(__m64 *)(s + 48);
- __m64 v7 = *(__m64 *)(s + 56);
- *(__m64 *)(d + 0) = v0;
- *(__m64 *)(d + 8) = v1;
- *(__m64 *)(d + 16) = v2;
- *(__m64 *)(d + 24) = v3;
- *(__m64 *)(d + 32) = v4;
- *(__m64 *)(d + 40) = v5;
- *(__m64 *)(d + 48) = v6;
- *(__m64 *)(d + 56) = v7;
-#endif
-
- w -= 64;
- s += 64;
- d += 64;
- }
- while (w >= 4)
- {
- *(uint32_t *)d = *(uint32_t *)s;
-
- w -= 4;
- s += 4;
- d += 4;
- }
- if (w >= 2)
- {
- *(uint16_t *)d = *(uint16_t *)s;
- w -= 2;
- s += 2;
- d += 2;
- }
- }
-
- _mm_empty ();
-
- return TRUE;
-}
-
-static void
-mmx_composite_copy_area (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- pixman_blt_mmx (src_image->bits.bits,
- dst_image->bits.bits,
- src_image->bits.rowstride,
- dst_image->bits.rowstride,
- PIXMAN_FORMAT_BPP (src_image->bits.format),
- PIXMAN_FORMAT_BPP (dst_image->bits.format),
- src_x, src_y, dest_x, dest_y, width, height);
-}
-
-#if 0
-static void
-mmx_composite_over_x888_8_8888 (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src_image,
- pixman_image_t * mask_image,
- pixman_image_t * dst_image,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height)
-{
- uint32_t *src, *src_line;
- uint32_t *dst, *dst_line;
- uint8_t *mask, *mask_line;
- int src_stride, mask_stride, dst_stride;
- int32_t w;
-
- PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1);
- PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1);
- PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
-
- while (height--)
- {
- src = src_line;
- src_line += src_stride;
- dst = dst_line;
- dst_line += dst_stride;
- mask = mask_line;
- mask_line += mask_stride;
-
- w = width;
-
- while (w--)
- {
- uint64_t m = *mask;
-
- if (m)
- {
- __m64 s = load8888 (*src | 0xff000000);
-
- if (m == 0xff)
- {
- *dst = store8888 (s);
- }
- else
- {
- __m64 sa = expand_alpha (s);
- __m64 vm = expand_alpha_rev (to_m64 (m));
- __m64 vdest = in_over (s, sa, vm, load8888 (*dst));
-
- *dst = store8888 (vdest);
- }
- }
-
- mask++;
- dst++;
- src++;
- }
- }
-
- _mm_empty ();
-}
-#endif
-
-static const pixman_fast_path_t mmx_fast_paths[] =
-{
- PIXMAN_STD_FAST_PATH (OVER, solid, a8, r5g6b5, mmx_composite_over_n_8_0565 ),
- PIXMAN_STD_FAST_PATH (OVER, solid, a8, b5g6r5, mmx_composite_over_n_8_0565 ),
- PIXMAN_STD_FAST_PATH (OVER, solid, a8, a8r8g8b8, mmx_composite_over_n_8_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, solid, a8, x8r8g8b8, mmx_composite_over_n_8_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, solid, a8, a8b8g8r8, mmx_composite_over_n_8_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, solid, a8, x8b8g8r8, mmx_composite_over_n_8_8888 ),
- PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8r8g8b8, a8r8g8b8, mmx_composite_over_n_8888_8888_ca ),
- PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8r8g8b8, x8r8g8b8, mmx_composite_over_n_8888_8888_ca ),
- PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8r8g8b8, r5g6b5, mmx_composite_over_n_8888_0565_ca ),
- PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8b8g8r8, a8b8g8r8, mmx_composite_over_n_8888_8888_ca ),
- PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8b8g8r8, x8b8g8r8, mmx_composite_over_n_8888_8888_ca ),
- PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8b8g8r8, b5g6r5, mmx_composite_over_n_8888_0565_ca ),
- PIXMAN_STD_FAST_PATH (OVER, pixbuf, pixbuf, a8r8g8b8, mmx_composite_over_pixbuf_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, pixbuf, pixbuf, x8r8g8b8, mmx_composite_over_pixbuf_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, pixbuf, pixbuf, r5g6b5, mmx_composite_over_pixbuf_0565 ),
- PIXMAN_STD_FAST_PATH (OVER, rpixbuf, rpixbuf, a8b8g8r8, mmx_composite_over_pixbuf_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, rpixbuf, rpixbuf, x8b8g8r8, mmx_composite_over_pixbuf_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, rpixbuf, rpixbuf, b5g6r5, mmx_composite_over_pixbuf_0565 ),
- PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, solid, a8r8g8b8, mmx_composite_over_x888_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, solid, x8r8g8b8, mmx_composite_over_x888_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, x8b8g8r8, solid, a8b8g8r8, mmx_composite_over_x888_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, x8b8g8r8, solid, x8b8g8r8, mmx_composite_over_x888_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, solid, a8r8g8b8, mmx_composite_over_8888_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, solid, x8r8g8b8, mmx_composite_over_8888_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, solid, a8b8g8r8, mmx_composite_over_8888_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, solid, x8b8g8r8, mmx_composite_over_8888_n_8888 ),
-#if 0
- /* FIXME: This code is commented out since it's apparently
- * not actually faster than the generic code.
- */
- PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, a8, x8r8g8b8, mmx_composite_over_x888_8_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, a8, a8r8g8b8, mmx_composite_over_x888_8_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, x8b8r8g8, a8, x8b8g8r8, mmx_composite_over_x888_8_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, x8b8r8g8, a8, a8r8g8b8, mmx_composite_over_x888_8_8888 ),
-#endif
- PIXMAN_STD_FAST_PATH (OVER, solid, null, a8r8g8b8, mmx_composite_over_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, solid, null, x8r8g8b8, mmx_composite_over_n_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, solid, null, r5g6b5, mmx_composite_over_n_0565 ),
- PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, null, x8r8g8b8, mmx_composite_copy_area ),
- PIXMAN_STD_FAST_PATH (OVER, x8b8g8r8, null, x8b8g8r8, mmx_composite_copy_area ),
-
- PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, null, a8r8g8b8, mmx_composite_over_8888_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, null, x8r8g8b8, mmx_composite_over_8888_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, null, r5g6b5, mmx_composite_over_8888_0565 ),
- PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, null, a8b8g8r8, mmx_composite_over_8888_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, null, x8b8g8r8, mmx_composite_over_8888_8888 ),
- PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, null, b5g6r5, mmx_composite_over_8888_0565 ),
-
- PIXMAN_STD_FAST_PATH (ADD, a8r8g8b8, null, a8r8g8b8, mmx_composite_add_8888_8888 ),
- PIXMAN_STD_FAST_PATH (ADD, a8b8g8r8, null, a8b8g8r8, mmx_composite_add_8888_8888 ),
- PIXMAN_STD_FAST_PATH (ADD, a8, null, a8, mmx_composite_add_8_8 ),
- PIXMAN_STD_FAST_PATH (ADD, solid, a8, a8, mmx_composite_add_n_8_8 ),
-
- PIXMAN_STD_FAST_PATH (SRC, solid, a8, a8r8g8b8, mmx_composite_src_n_8_8888 ),
- PIXMAN_STD_FAST_PATH (SRC, solid, a8, x8r8g8b8, mmx_composite_src_n_8_8888 ),
- PIXMAN_STD_FAST_PATH (SRC, solid, a8, a8b8g8r8, mmx_composite_src_n_8_8888 ),
- PIXMAN_STD_FAST_PATH (SRC, solid, a8, x8b8g8r8, mmx_composite_src_n_8_8888 ),
- PIXMAN_STD_FAST_PATH (SRC, a8r8g8b8, null, a8r8g8b8, mmx_composite_copy_area ),
- PIXMAN_STD_FAST_PATH (SRC, a8b8g8r8, null, a8b8g8r8, mmx_composite_copy_area ),
- PIXMAN_STD_FAST_PATH (SRC, a8r8g8b8, null, x8r8g8b8, mmx_composite_copy_area ),
- PIXMAN_STD_FAST_PATH (SRC, a8b8g8r8, null, x8b8g8r8, mmx_composite_copy_area ),
- PIXMAN_STD_FAST_PATH (SRC, x8r8g8b8, null, x8r8g8b8, mmx_composite_copy_area ),
- PIXMAN_STD_FAST_PATH (SRC, x8b8g8r8, null, x8b8g8r8, mmx_composite_copy_area ),
- PIXMAN_STD_FAST_PATH (SRC, r5g6b5, null, r5g6b5, mmx_composite_copy_area ),
- PIXMAN_STD_FAST_PATH (SRC, b5g6r5, null, b5g6r5, mmx_composite_copy_area ),
-
- PIXMAN_STD_FAST_PATH (IN, a8, null, a8, mmx_composite_in_8_8 ),
- PIXMAN_STD_FAST_PATH (IN, solid, a8, a8, mmx_composite_in_n_8_8 ),
-
- { PIXMAN_OP_NONE },
-};
-
-static pixman_bool_t
-mmx_blt (pixman_implementation_t *imp,
- uint32_t * src_bits,
- uint32_t * dst_bits,
- int src_stride,
- int dst_stride,
- int src_bpp,
- int dst_bpp,
- int src_x,
- int src_y,
- int dst_x,
- int dst_y,
- int width,
- int height)
-{
- if (!pixman_blt_mmx (
- src_bits, dst_bits, src_stride, dst_stride, src_bpp, dst_bpp,
- src_x, src_y, dst_x, dst_y, width, height))
-
- {
- return _pixman_implementation_blt (
- imp->delegate,
- src_bits, dst_bits, src_stride, dst_stride, src_bpp, dst_bpp,
- src_x, src_y, dst_x, dst_y, width, height);
- }
-
- return TRUE;
-}
-
-static pixman_bool_t
-mmx_fill (pixman_implementation_t *imp,
- uint32_t * bits,
- int stride,
- int bpp,
- int x,
- int y,
- int width,
- int height,
- uint32_t xor)
-{
- if (!pixman_fill_mmx (bits, stride, bpp, x, y, width, height, xor))
- {
- return _pixman_implementation_fill (
- imp->delegate, bits, stride, bpp, x, y, width, height, xor);
- }
-
- return TRUE;
-}
-
-pixman_implementation_t *
-_pixman_implementation_create_mmx (void)
-{
- pixman_implementation_t *general = _pixman_implementation_create_fast_path ();
- pixman_implementation_t *imp = _pixman_implementation_create (general, mmx_fast_paths);
-
- imp->combine_32[PIXMAN_OP_OVER] = mmx_combine_over_u;
- imp->combine_32[PIXMAN_OP_OVER_REVERSE] = mmx_combine_over_reverse_u;
- imp->combine_32[PIXMAN_OP_IN] = mmx_combine_in_u;
- imp->combine_32[PIXMAN_OP_IN_REVERSE] = mmx_combine_in_reverse_u;
- imp->combine_32[PIXMAN_OP_OUT] = mmx_combine_out_u;
- imp->combine_32[PIXMAN_OP_OUT_REVERSE] = mmx_combine_out_reverse_u;
- imp->combine_32[PIXMAN_OP_ATOP] = mmx_combine_atop_u;
- imp->combine_32[PIXMAN_OP_ATOP_REVERSE] = mmx_combine_atop_reverse_u;
- imp->combine_32[PIXMAN_OP_XOR] = mmx_combine_xor_u;
- imp->combine_32[PIXMAN_OP_ADD] = mmx_combine_add_u;
- imp->combine_32[PIXMAN_OP_SATURATE] = mmx_combine_saturate_u;
-
- imp->combine_32_ca[PIXMAN_OP_SRC] = mmx_combine_src_ca;
- imp->combine_32_ca[PIXMAN_OP_OVER] = mmx_combine_over_ca;
- imp->combine_32_ca[PIXMAN_OP_OVER_REVERSE] = mmx_combine_over_reverse_ca;
- imp->combine_32_ca[PIXMAN_OP_IN] = mmx_combine_in_ca;
- imp->combine_32_ca[PIXMAN_OP_IN_REVERSE] = mmx_combine_in_reverse_ca;
- imp->combine_32_ca[PIXMAN_OP_OUT] = mmx_combine_out_ca;
- imp->combine_32_ca[PIXMAN_OP_OUT_REVERSE] = mmx_combine_out_reverse_ca;
- imp->combine_32_ca[PIXMAN_OP_ATOP] = mmx_combine_atop_ca;
- imp->combine_32_ca[PIXMAN_OP_ATOP_REVERSE] = mmx_combine_atop_reverse_ca;
- imp->combine_32_ca[PIXMAN_OP_XOR] = mmx_combine_xor_ca;
- imp->combine_32_ca[PIXMAN_OP_ADD] = mmx_combine_add_ca;
-
- imp->blt = mmx_blt;
- imp->fill = mmx_fill;
-
- return imp;
-}
-
-#endif /* USE_MMX */
+/* + * Copyright © 2004, 2005 Red Hat, Inc. + * Copyright © 2004 Nicholas Miell + * Copyright © 2005 Trolltech AS + * + * Permission to use, copy, modify, distribute, and sell this software and its + * documentation for any purpose is hereby granted without fee, provided that + * the above copyright notice appear in all copies and that both that + * copyright notice and this permission notice appear in supporting + * documentation, and that the name of Red Hat not be used in advertising or + * publicity pertaining to distribution of the software without specific, + * written prior permission. Red Hat makes no representations about the + * suitability of this software for any purpose. It is provided "as is" + * without express or implied warranty. + * + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS + * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND + * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY + * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN + * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING + * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS + * SOFTWARE. + * + * Author: Søren Sandmann (sandmann@redhat.com) + * Minor Improvements: Nicholas Miell (nmiell@gmail.com) + * MMX code paths for fbcompose.c by Lars Knoll (lars@trolltech.com) + * + * Based on work by Owen Taylor + */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#ifdef USE_MMX + +#include <mmintrin.h> +#include "pixman-private.h" +#include "pixman-combine32.h" + +#define no_vERBOSE + +#ifdef VERBOSE +#define CHECKPOINT() error_f ("at %s %d\n", __FUNCTION__, __LINE__) +#else +#define CHECKPOINT() +#endif + +/* Notes about writing mmx code + * + * give memory operands as the second operand. If you give it as the + * first, gcc will first load it into a register, then use that + * register + * + * ie. use + * + * _mm_mullo_pi16 (x, mmx_constant); + * + * not + * + * _mm_mullo_pi16 (mmx_constant, x); + * + * Also try to minimize dependencies. i.e. when you need a value, try + * to calculate it from a value that was calculated as early as + * possible. + */ + +/* --------------- MMX primitives ------------------------------------- */ + +#ifdef __GNUC__ +typedef uint64_t mmxdatafield; +#else +typedef __m64 mmxdatafield; +/* If __m64 is defined as a struct or union, define M64_MEMBER to be the + name of the member used to access the data */ +# ifdef _MSC_VER +# define M64_MEMBER m64_u64 +# elif defined(__SUNPRO_C) +# define M64_MEMBER l_ +# endif +#endif + +typedef struct +{ + mmxdatafield mmx_4x00ff; + mmxdatafield mmx_4x0080; + mmxdatafield mmx_565_rgb; + mmxdatafield mmx_565_unpack_multiplier; + mmxdatafield mmx_565_r; + mmxdatafield mmx_565_g; + mmxdatafield mmx_565_b; + mmxdatafield mmx_mask_0; + mmxdatafield mmx_mask_1; + mmxdatafield mmx_mask_2; + mmxdatafield mmx_mask_3; + mmxdatafield mmx_full_alpha; + mmxdatafield mmx_ffff0000ffff0000; + mmxdatafield mmx_0000ffff00000000; + mmxdatafield mmx_000000000000ffff; +} mmx_data_t; + +#if defined(_MSC_VER) +# define MMXDATA_INIT(field, val) { val ## UI64 } +#elif defined(M64_MEMBER) /* __m64 is a struct, not an integral type */ +# define MMXDATA_INIT(field, val) field = { val ## ULL } +#else /* __m64 is an integral type */ +# define MMXDATA_INIT(field, val) field = val ## ULL +#endif + +static const mmx_data_t c = +{ + MMXDATA_INIT (.mmx_4x00ff, 0x00ff00ff00ff00ff), + MMXDATA_INIT (.mmx_4x0080, 0x0080008000800080), + MMXDATA_INIT (.mmx_565_rgb, 0x000001f0003f001f), + MMXDATA_INIT (.mmx_565_unpack_multiplier, 0x0000008404100840), + MMXDATA_INIT (.mmx_565_r, 0x000000f800000000), + MMXDATA_INIT (.mmx_565_g, 0x0000000000fc0000), + MMXDATA_INIT (.mmx_565_b, 0x00000000000000f8), + MMXDATA_INIT (.mmx_mask_0, 0xffffffffffff0000), + MMXDATA_INIT (.mmx_mask_1, 0xffffffff0000ffff), + MMXDATA_INIT (.mmx_mask_2, 0xffff0000ffffffff), + MMXDATA_INIT (.mmx_mask_3, 0x0000ffffffffffff), + MMXDATA_INIT (.mmx_full_alpha, 0x00ff000000000000), + MMXDATA_INIT (.mmx_ffff0000ffff0000, 0xffff0000ffff0000), + MMXDATA_INIT (.mmx_0000ffff00000000, 0x0000ffff00000000), + MMXDATA_INIT (.mmx_000000000000ffff, 0x000000000000ffff), +}; + +#ifdef __GNUC__ +# ifdef __ICC +# define MC(x) to_m64 (c.mmx_ ## x) +# else +# define MC(x) ((__m64)c.mmx_ ## x) +# endif +#else +# define MC(x) c.mmx_ ## x +#endif + +static force_inline __m64 +to_m64 (uint64_t x) +{ +#ifdef __ICC + return _mm_cvtsi64_m64 (x); +#elif defined M64_MEMBER /* __m64 is a struct, not an integral type */ + __m64 res; + + res.M64_MEMBER = x; + return res; +#else /* __m64 is an integral type */ + return (__m64)x; +#endif +} + +static force_inline uint64_t +to_uint64 (__m64 x) +{ +#ifdef __ICC + return _mm_cvtm64_si64 (x); +#elif defined M64_MEMBER /* __m64 is a struct, not an integral type */ + uint64_t res = x.M64_MEMBER; + return res; +#else /* __m64 is an integral type */ + return (uint64_t)x; +#endif +} + +static force_inline __m64 +shift (__m64 v, + int s) +{ + if (s > 0) + return _mm_slli_si64 (v, s); + else if (s < 0) + return _mm_srli_si64 (v, -s); + else + return v; +} + +static force_inline __m64 +negate (__m64 mask) +{ + return _mm_xor_si64 (mask, MC (4x00ff)); +} + +static force_inline __m64 +pix_multiply (__m64 a, __m64 b) +{ + __m64 res; + + res = _mm_mullo_pi16 (a, b); + res = _mm_adds_pu16 (res, MC (4x0080)); + res = _mm_adds_pu16 (res, _mm_srli_pi16 (res, 8)); + res = _mm_srli_pi16 (res, 8); + + return res; +} + +static force_inline __m64 +pix_add (__m64 a, __m64 b) +{ + return _mm_adds_pu8 (a, b); +} + +static force_inline __m64 +expand_alpha (__m64 pixel) +{ + __m64 t1, t2; + + t1 = shift (pixel, -48); + t2 = shift (t1, 16); + t1 = _mm_or_si64 (t1, t2); + t2 = shift (t1, 32); + t1 = _mm_or_si64 (t1, t2); + + return t1; +} + +static force_inline __m64 +expand_alpha_rev (__m64 pixel) +{ + __m64 t1, t2; + + /* move alpha to low 16 bits and zero the rest */ + t1 = shift (pixel, 48); + t1 = shift (t1, -48); + + t2 = shift (t1, 16); + t1 = _mm_or_si64 (t1, t2); + t2 = shift (t1, 32); + t1 = _mm_or_si64 (t1, t2); + + return t1; +} + +static force_inline __m64 +invert_colors (__m64 pixel) +{ + __m64 x, y, z; + + x = y = z = pixel; + + x = _mm_and_si64 (x, MC (ffff0000ffff0000)); + y = _mm_and_si64 (y, MC (000000000000ffff)); + z = _mm_and_si64 (z, MC (0000ffff00000000)); + + y = shift (y, 32); + z = shift (z, -32); + + x = _mm_or_si64 (x, y); + x = _mm_or_si64 (x, z); + + return x; +} + +static force_inline __m64 +over (__m64 src, + __m64 srca, + __m64 dest) +{ + return _mm_adds_pu8 (src, pix_multiply (dest, negate (srca))); +} + +static force_inline __m64 +over_rev_non_pre (__m64 src, __m64 dest) +{ + __m64 srca = expand_alpha (src); + __m64 srcfaaa = _mm_or_si64 (srca, MC (full_alpha)); + + return over (pix_multiply (invert_colors (src), srcfaaa), srca, dest); +} + +static force_inline __m64 +in (__m64 src, __m64 mask) +{ + return pix_multiply (src, mask); +} + +static force_inline __m64 +in_over_full_src_alpha (__m64 src, __m64 mask, __m64 dest) +{ + src = _mm_or_si64 (src, MC (full_alpha)); + + return over (in (src, mask), mask, dest); +} + +#ifndef _MSC_VER +static force_inline __m64 +in_over (__m64 src, __m64 srca, __m64 mask, __m64 dest) +{ + return over (in (src, mask), pix_multiply (srca, mask), dest); +} + +#else + +#define in_over(src, srca, mask, dest) \ + over (in (src, mask), pix_multiply (srca, mask), dest) + +#endif + +static force_inline __m64 +load8888 (uint32_t v) +{ + return _mm_unpacklo_pi8 (_mm_cvtsi32_si64 (v), _mm_setzero_si64 ()); +} + +static force_inline __m64 +pack8888 (__m64 lo, __m64 hi) +{ + return _mm_packs_pu16 (lo, hi); +} + +static force_inline uint32_t +store8888 (__m64 v) +{ + return _mm_cvtsi64_si32 (pack8888 (v, _mm_setzero_si64 ())); +} + +/* Expand 16 bits positioned at @pos (0-3) of a mmx register into + * + * 00RR00GG00BB + * + * --- Expanding 565 in the low word --- + * + * m = (m << (32 - 3)) | (m << (16 - 5)) | m; + * m = m & (01f0003f001f); + * m = m * (008404100840); + * m = m >> 8; + * + * Note the trick here - the top word is shifted by another nibble to + * avoid it bumping into the middle word + */ +static force_inline __m64 +expand565 (__m64 pixel, int pos) +{ + __m64 p = pixel; + __m64 t1, t2; + + /* move pixel to low 16 bit and zero the rest */ + p = shift (shift (p, (3 - pos) * 16), -48); + + t1 = shift (p, 36 - 11); + t2 = shift (p, 16 - 5); + + p = _mm_or_si64 (t1, p); + p = _mm_or_si64 (t2, p); + p = _mm_and_si64 (p, MC (565_rgb)); + + pixel = _mm_mullo_pi16 (p, MC (565_unpack_multiplier)); + return _mm_srli_pi16 (pixel, 8); +} + +static force_inline __m64 +expand8888 (__m64 in, int pos) +{ + if (pos == 0) + return _mm_unpacklo_pi8 (in, _mm_setzero_si64 ()); + else + return _mm_unpackhi_pi8 (in, _mm_setzero_si64 ()); +} + +static force_inline __m64 +expandx888 (__m64 in, int pos) +{ + return _mm_or_si64 (expand8888 (in, pos), MC (full_alpha)); +} + +static force_inline __m64 +pack_565 (__m64 pixel, __m64 target, int pos) +{ + __m64 p = pixel; + __m64 t = target; + __m64 r, g, b; + + r = _mm_and_si64 (p, MC (565_r)); + g = _mm_and_si64 (p, MC (565_g)); + b = _mm_and_si64 (p, MC (565_b)); + + r = shift (r, -(32 - 8) + pos * 16); + g = shift (g, -(16 - 3) + pos * 16); + b = shift (b, -(0 + 3) + pos * 16); + + if (pos == 0) + t = _mm_and_si64 (t, MC (mask_0)); + else if (pos == 1) + t = _mm_and_si64 (t, MC (mask_1)); + else if (pos == 2) + t = _mm_and_si64 (t, MC (mask_2)); + else if (pos == 3) + t = _mm_and_si64 (t, MC (mask_3)); + + p = _mm_or_si64 (r, t); + p = _mm_or_si64 (g, p); + + return _mm_or_si64 (b, p); +} + +#ifndef _MSC_VER + +static force_inline __m64 +pix_add_mul (__m64 x, __m64 a, __m64 y, __m64 b) +{ + x = pix_multiply (x, a); + y = pix_multiply (y, b); + + return pix_add (x, y); +} + +#else + +#define pix_add_mul(x, a, y, b) \ + ( x = pix_multiply (x, a), \ + y = pix_multiply (y, a), \ + pix_add (x, y) ) + +#endif + +/* --------------- MMX code patch for fbcompose.c --------------------- */ + +static force_inline uint32_t +combine (const uint32_t *src, const uint32_t *mask) +{ + uint32_t ssrc = *src; + + if (mask) + { + __m64 m = load8888 (*mask); + __m64 s = load8888 (ssrc); + + m = expand_alpha (m); + s = pix_multiply (s, m); + + ssrc = store8888 (s); + } + + return ssrc; +} + +static void +mmx_combine_over_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + uint32_t ssrc = combine (src, mask); + uint32_t a = ssrc >> 24; + + if (a == 0xff) + { + *dest = ssrc; + } + else if (ssrc) + { + __m64 s, sa; + s = load8888 (ssrc); + sa = expand_alpha (s); + *dest = store8888 (over (s, sa, load8888 (*dest))); + } + + ++dest; + ++src; + if (mask) + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_over_reverse_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + __m64 d, da; + uint32_t s = combine (src, mask); + + d = load8888 (*dest); + da = expand_alpha (d); + *dest = store8888 (over (d, da, load8888 (s))); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_in_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + __m64 x, a; + + x = load8888 (combine (src, mask)); + a = load8888 (*dest); + a = expand_alpha (a); + x = pix_multiply (x, a); + + *dest = store8888 (x); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_in_reverse_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + __m64 x, a; + + x = load8888 (*dest); + a = load8888 (combine (src, mask)); + a = expand_alpha (a); + x = pix_multiply (x, a); + *dest = store8888 (x); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_out_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + __m64 x, a; + + x = load8888 (combine (src, mask)); + a = load8888 (*dest); + a = expand_alpha (a); + a = negate (a); + x = pix_multiply (x, a); + *dest = store8888 (x); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_out_reverse_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + __m64 x, a; + + x = load8888 (*dest); + a = load8888 (combine (src, mask)); + a = expand_alpha (a); + a = negate (a); + x = pix_multiply (x, a); + + *dest = store8888 (x); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_atop_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + __m64 s, da, d, sia; + + s = load8888 (combine (src, mask)); + d = load8888 (*dest); + sia = expand_alpha (s); + sia = negate (sia); + da = expand_alpha (d); + s = pix_add_mul (s, da, d, sia); + *dest = store8888 (s); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_atop_reverse_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end; + + end = dest + width; + + while (dest < end) + { + __m64 s, dia, d, sa; + + s = load8888 (combine (src, mask)); + d = load8888 (*dest); + sa = expand_alpha (s); + dia = expand_alpha (d); + dia = negate (dia); + s = pix_add_mul (s, dia, d, sa); + *dest = store8888 (s); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_xor_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + __m64 s, dia, d, sia; + + s = load8888 (combine (src, mask)); + d = load8888 (*dest); + sia = expand_alpha (s); + dia = expand_alpha (d); + sia = negate (sia); + dia = negate (dia); + s = pix_add_mul (s, dia, d, sia); + *dest = store8888 (s); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_add_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + __m64 s, d; + + s = load8888 (combine (src, mask)); + d = load8888 (*dest); + s = pix_add (s, d); + *dest = store8888 (s); + + ++dest; + ++src; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_saturate_u (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = dest + width; + + while (dest < end) + { + uint32_t s = combine (src, mask); + uint32_t d = *dest; + __m64 ms = load8888 (s); + __m64 md = load8888 (d); + uint32_t sa = s >> 24; + uint32_t da = ~d >> 24; + + if (sa > da) + { + __m64 msa = load8888 (DIV_UN8 (da, sa) << 24); + msa = expand_alpha (msa); + ms = pix_multiply (ms, msa); + } + + md = pix_add (md, ms); + *dest = store8888 (md); + + ++src; + ++dest; + if (mask) + mask++; + } + _mm_empty (); +} + +static void +mmx_combine_src_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + + s = pix_multiply (s, a); + *dest = store8888 (s); + + ++src; + ++mask; + ++dest; + } + _mm_empty (); +} + +static void +mmx_combine_over_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 sa = expand_alpha (s); + + *dest = store8888 (in_over (s, sa, a, d)); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_over_reverse_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 da = expand_alpha (d); + + *dest = store8888 (over (d, da, in (s, a))); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_in_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 da = expand_alpha (d); + + s = pix_multiply (s, a); + s = pix_multiply (s, da); + *dest = store8888 (s); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_in_reverse_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 sa = expand_alpha (s); + + a = pix_multiply (a, sa); + d = pix_multiply (d, a); + *dest = store8888 (d); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_out_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 da = expand_alpha (d); + + da = negate (da); + s = pix_multiply (s, a); + s = pix_multiply (s, da); + *dest = store8888 (s); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_out_reverse_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 sa = expand_alpha (s); + + a = pix_multiply (a, sa); + a = negate (a); + d = pix_multiply (d, a); + *dest = store8888 (d); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_atop_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 da = expand_alpha (d); + __m64 sa = expand_alpha (s); + + s = pix_multiply (s, a); + a = pix_multiply (a, sa); + a = negate (a); + d = pix_add_mul (d, a, s, da); + *dest = store8888 (d); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_atop_reverse_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 da = expand_alpha (d); + __m64 sa = expand_alpha (s); + + s = pix_multiply (s, a); + a = pix_multiply (a, sa); + da = negate (da); + d = pix_add_mul (d, a, s, da); + *dest = store8888 (d); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_xor_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + __m64 da = expand_alpha (d); + __m64 sa = expand_alpha (s); + + s = pix_multiply (s, a); + a = pix_multiply (a, sa); + da = negate (da); + a = negate (a); + d = pix_add_mul (d, a, s, da); + *dest = store8888 (d); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +static void +mmx_combine_add_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width) +{ + const uint32_t *end = src + width; + + while (src < end) + { + __m64 a = load8888 (*mask); + __m64 s = load8888 (*src); + __m64 d = load8888 (*dest); + + s = pix_multiply (s, a); + d = pix_add (s, d); + *dest = store8888 (d); + + ++src; + ++dest; + ++mask; + } + _mm_empty (); +} + +/* ------------- MMX code paths called from fbpict.c -------------------- */ + +static void +mmx_composite_over_n_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t src; + uint32_t *dst_line, *dst; + int32_t w; + int dst_stride; + __m64 vsrc, vsrca; + + CHECKPOINT (); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + if (src == 0) + return; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + w = width; + + CHECKPOINT (); + + while (w && (unsigned long)dst & 7) + { + *dst = store8888 (over (vsrc, vsrca, load8888 (*dst))); + + w--; + dst++; + } + + while (w >= 2) + { + __m64 vdest; + __m64 dest0, dest1; + + vdest = *(__m64 *)dst; + + dest0 = over (vsrc, vsrca, expand8888 (vdest, 0)); + dest1 = over (vsrc, vsrca, expand8888 (vdest, 1)); + + *(__m64 *)dst = pack8888 (dest0, dest1); + + dst += 2; + w -= 2; + } + + CHECKPOINT (); + + while (w) + { + *dst = store8888 (over (vsrc, vsrca, load8888 (*dst))); + + w--; + dst++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_n_0565 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t src; + uint16_t *dst_line, *dst; + int32_t w; + int dst_stride; + __m64 vsrc, vsrca; + + CHECKPOINT (); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + if (src == 0) + return; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1); + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + w = width; + + CHECKPOINT (); + + while (w && (unsigned long)dst & 7) + { + uint64_t d = *dst; + __m64 vdest = expand565 (to_m64 (d), 0); + + vdest = pack_565 (over (vsrc, vsrca, vdest), vdest, 0); + *dst = to_uint64 (vdest); + + w--; + dst++; + } + + while (w >= 4) + { + __m64 vdest; + + vdest = *(__m64 *)dst; + + vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 0)), vdest, 0); + vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 1)), vdest, 1); + vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 2)), vdest, 2); + vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 3)), vdest, 3); + + *(__m64 *)dst = vdest; + + dst += 4; + w -= 4; + } + + CHECKPOINT (); + + while (w) + { + uint64_t d = *dst; + __m64 vdest = expand565 (to_m64 (d), 0); + + vdest = pack_565 (over (vsrc, vsrca, vdest), vdest, 0); + *dst = to_uint64 (vdest); + + w--; + dst++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_n_8888_8888_ca (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t src, srca; + uint32_t *dst_line; + uint32_t *mask_line; + int dst_stride, mask_stride; + __m64 vsrc, vsrca; + + CHECKPOINT (); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + srca = src >> 24; + if (src == 0) + return; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint32_t, mask_stride, mask_line, 1); + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + while (height--) + { + int twidth = width; + uint32_t *p = (uint32_t *)mask_line; + uint32_t *q = (uint32_t *)dst_line; + + while (twidth && (unsigned long)q & 7) + { + uint32_t m = *(uint32_t *)p; + + if (m) + { + __m64 vdest = load8888 (*q); + vdest = in_over (vsrc, vsrca, load8888 (m), vdest); + *q = store8888 (vdest); + } + + twidth--; + p++; + q++; + } + + while (twidth >= 2) + { + uint32_t m0, m1; + m0 = *p; + m1 = *(p + 1); + + if (m0 | m1) + { + __m64 dest0, dest1; + __m64 vdest = *(__m64 *)q; + + dest0 = in_over (vsrc, vsrca, load8888 (m0), + expand8888 (vdest, 0)); + dest1 = in_over (vsrc, vsrca, load8888 (m1), + expand8888 (vdest, 1)); + + *(__m64 *)q = pack8888 (dest0, dest1); + } + + p += 2; + q += 2; + twidth -= 2; + } + + while (twidth) + { + uint32_t m = *(uint32_t *)p; + + if (m) + { + __m64 vdest = load8888 (*q); + vdest = in_over (vsrc, vsrca, load8888 (m), vdest); + *q = store8888 (vdest); + } + + twidth--; + p++; + q++; + } + + dst_line += dst_stride; + mask_line += mask_stride; + } + + _mm_empty (); +} + +static void +mmx_composite_over_8888_n_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t *dst_line, *dst; + uint32_t *src_line, *src; + uint32_t mask; + __m64 vmask; + int dst_stride, src_stride; + int32_t w; + __m64 srca; + + CHECKPOINT (); + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); + + mask = _pixman_image_get_solid (imp, mask_image, dst_image->bits.format); + mask &= 0xff000000; + mask = mask | mask >> 8 | mask >> 16 | mask >> 24; + vmask = load8888 (mask); + srca = MC (4x00ff); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + while (w && (unsigned long)dst & 7) + { + __m64 s = load8888 (*src); + __m64 d = load8888 (*dst); + + *dst = store8888 (in_over (s, expand_alpha (s), vmask, d)); + + w--; + dst++; + src++; + } + + while (w >= 2) + { + __m64 vs = *(__m64 *)src; + __m64 vd = *(__m64 *)dst; + __m64 vsrc0 = expand8888 (vs, 0); + __m64 vsrc1 = expand8888 (vs, 1); + + *(__m64 *)dst = pack8888 ( + in_over (vsrc0, expand_alpha (vsrc0), vmask, expand8888 (vd, 0)), + in_over (vsrc1, expand_alpha (vsrc1), vmask, expand8888 (vd, 1))); + + w -= 2; + dst += 2; + src += 2; + } + + while (w) + { + __m64 s = load8888 (*src); + __m64 d = load8888 (*dst); + + *dst = store8888 (in_over (s, expand_alpha (s), vmask, d)); + + w--; + dst++; + src++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_x888_n_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t *dst_line, *dst; + uint32_t *src_line, *src; + uint32_t mask; + __m64 vmask; + int dst_stride, src_stride; + int32_t w; + __m64 srca; + + CHECKPOINT (); + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); + mask = _pixman_image_get_solid (imp, mask_image, dst_image->bits.format); + + mask &= 0xff000000; + mask = mask | mask >> 8 | mask >> 16 | mask >> 24; + vmask = load8888 (mask); + srca = MC (4x00ff); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + while (w && (unsigned long)dst & 7) + { + __m64 s = load8888 (*src | 0xff000000); + __m64 d = load8888 (*dst); + + *dst = store8888 (in_over (s, srca, vmask, d)); + + w--; + dst++; + src++; + } + + while (w >= 16) + { + __m64 vd0 = *(__m64 *)(dst + 0); + __m64 vd1 = *(__m64 *)(dst + 2); + __m64 vd2 = *(__m64 *)(dst + 4); + __m64 vd3 = *(__m64 *)(dst + 6); + __m64 vd4 = *(__m64 *)(dst + 8); + __m64 vd5 = *(__m64 *)(dst + 10); + __m64 vd6 = *(__m64 *)(dst + 12); + __m64 vd7 = *(__m64 *)(dst + 14); + + __m64 vs0 = *(__m64 *)(src + 0); + __m64 vs1 = *(__m64 *)(src + 2); + __m64 vs2 = *(__m64 *)(src + 4); + __m64 vs3 = *(__m64 *)(src + 6); + __m64 vs4 = *(__m64 *)(src + 8); + __m64 vs5 = *(__m64 *)(src + 10); + __m64 vs6 = *(__m64 *)(src + 12); + __m64 vs7 = *(__m64 *)(src + 14); + + vd0 = pack8888 ( + in_over (expandx888 (vs0, 0), srca, vmask, expand8888 (vd0, 0)), + in_over (expandx888 (vs0, 1), srca, vmask, expand8888 (vd0, 1))); + + vd1 = pack8888 ( + in_over (expandx888 (vs1, 0), srca, vmask, expand8888 (vd1, 0)), + in_over (expandx888 (vs1, 1), srca, vmask, expand8888 (vd1, 1))); + + vd2 = pack8888 ( + in_over (expandx888 (vs2, 0), srca, vmask, expand8888 (vd2, 0)), + in_over (expandx888 (vs2, 1), srca, vmask, expand8888 (vd2, 1))); + + vd3 = pack8888 ( + in_over (expandx888 (vs3, 0), srca, vmask, expand8888 (vd3, 0)), + in_over (expandx888 (vs3, 1), srca, vmask, expand8888 (vd3, 1))); + + vd4 = pack8888 ( + in_over (expandx888 (vs4, 0), srca, vmask, expand8888 (vd4, 0)), + in_over (expandx888 (vs4, 1), srca, vmask, expand8888 (vd4, 1))); + + vd5 = pack8888 ( + in_over (expandx888 (vs5, 0), srca, vmask, expand8888 (vd5, 0)), + in_over (expandx888 (vs5, 1), srca, vmask, expand8888 (vd5, 1))); + + vd6 = pack8888 ( + in_over (expandx888 (vs6, 0), srca, vmask, expand8888 (vd6, 0)), + in_over (expandx888 (vs6, 1), srca, vmask, expand8888 (vd6, 1))); + + vd7 = pack8888 ( + in_over (expandx888 (vs7, 0), srca, vmask, expand8888 (vd7, 0)), + in_over (expandx888 (vs7, 1), srca, vmask, expand8888 (vd7, 1))); + + *(__m64 *)(dst + 0) = vd0; + *(__m64 *)(dst + 2) = vd1; + *(__m64 *)(dst + 4) = vd2; + *(__m64 *)(dst + 6) = vd3; + *(__m64 *)(dst + 8) = vd4; + *(__m64 *)(dst + 10) = vd5; + *(__m64 *)(dst + 12) = vd6; + *(__m64 *)(dst + 14) = vd7; + + w -= 16; + dst += 16; + src += 16; + } + + while (w) + { + __m64 s = load8888 (*src | 0xff000000); + __m64 d = load8888 (*dst); + + *dst = store8888 (in_over (s, srca, vmask, d)); + + w--; + dst++; + src++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_8888_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t *dst_line, *dst; + uint32_t *src_line, *src; + uint32_t s; + int dst_stride, src_stride; + uint8_t a; + int32_t w; + + CHECKPOINT (); + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + while (w--) + { + s = *src++; + a = s >> 24; + + if (a == 0xff) + { + *dst = s; + } + else if (s) + { + __m64 ms, sa; + ms = load8888 (s); + sa = expand_alpha (ms); + *dst = store8888 (over (ms, sa, load8888 (*dst))); + } + + dst++; + } + } + _mm_empty (); +} + +static void +mmx_composite_over_8888_0565 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint16_t *dst_line, *dst; + uint32_t *src_line, *src; + int dst_stride, src_stride; + int32_t w; + + CHECKPOINT (); + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); + +#if 0 + /* FIXME */ + assert (src_image->drawable == mask_image->drawable); +#endif + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + CHECKPOINT (); + + while (w && (unsigned long)dst & 7) + { + __m64 vsrc = load8888 (*src); + uint64_t d = *dst; + __m64 vdest = expand565 (to_m64 (d), 0); + + vdest = pack_565 ( + over (vsrc, expand_alpha (vsrc), vdest), vdest, 0); + + *dst = to_uint64 (vdest); + + w--; + dst++; + src++; + } + + CHECKPOINT (); + + while (w >= 4) + { + __m64 vsrc0, vsrc1, vsrc2, vsrc3; + __m64 vdest; + + vsrc0 = load8888 (*(src + 0)); + vsrc1 = load8888 (*(src + 1)); + vsrc2 = load8888 (*(src + 2)); + vsrc3 = load8888 (*(src + 3)); + + vdest = *(__m64 *)dst; + + vdest = pack_565 (over (vsrc0, expand_alpha (vsrc0), expand565 (vdest, 0)), vdest, 0); + vdest = pack_565 (over (vsrc1, expand_alpha (vsrc1), expand565 (vdest, 1)), vdest, 1); + vdest = pack_565 (over (vsrc2, expand_alpha (vsrc2), expand565 (vdest, 2)), vdest, 2); + vdest = pack_565 (over (vsrc3, expand_alpha (vsrc3), expand565 (vdest, 3)), vdest, 3); + + *(__m64 *)dst = vdest; + + w -= 4; + dst += 4; + src += 4; + } + + CHECKPOINT (); + + while (w) + { + __m64 vsrc = load8888 (*src); + uint64_t d = *dst; + __m64 vdest = expand565 (to_m64 (d), 0); + + vdest = pack_565 (over (vsrc, expand_alpha (vsrc), vdest), vdest, 0); + + *dst = to_uint64 (vdest); + + w--; + dst++; + src++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_n_8_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t src, srca; + uint32_t *dst_line, *dst; + uint8_t *mask_line, *mask; + int dst_stride, mask_stride; + int32_t w; + __m64 vsrc, vsrca; + uint64_t srcsrc; + + CHECKPOINT (); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + srca = src >> 24; + if (src == 0) + return; + + srcsrc = (uint64_t)src << 32 | src; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + mask = mask_line; + mask_line += mask_stride; + w = width; + + CHECKPOINT (); + + while (w && (unsigned long)dst & 7) + { + uint64_t m = *mask; + + if (m) + { + __m64 vdest = in_over (vsrc, vsrca, + expand_alpha_rev (to_m64 (m)), + load8888 (*dst)); + + *dst = store8888 (vdest); + } + + w--; + mask++; + dst++; + } + + CHECKPOINT (); + + while (w >= 2) + { + uint64_t m0, m1; + + m0 = *mask; + m1 = *(mask + 1); + + if (srca == 0xff && (m0 & m1) == 0xff) + { + *(uint64_t *)dst = srcsrc; + } + else if (m0 | m1) + { + __m64 vdest; + __m64 dest0, dest1; + + vdest = *(__m64 *)dst; + + dest0 = in_over (vsrc, vsrca, expand_alpha_rev (to_m64 (m0)), + expand8888 (vdest, 0)); + dest1 = in_over (vsrc, vsrca, expand_alpha_rev (to_m64 (m1)), + expand8888 (vdest, 1)); + + *(__m64 *)dst = pack8888 (dest0, dest1); + } + + mask += 2; + dst += 2; + w -= 2; + } + + CHECKPOINT (); + + while (w) + { + uint64_t m = *mask; + + if (m) + { + __m64 vdest = load8888 (*dst); + + vdest = in_over ( + vsrc, vsrca, expand_alpha_rev (to_m64 (m)), vdest); + *dst = store8888 (vdest); + } + + w--; + mask++; + dst++; + } + } + + _mm_empty (); +} + +pixman_bool_t +pixman_fill_mmx (uint32_t *bits, + int stride, + int bpp, + int x, + int y, + int width, + int height, + uint32_t xor) +{ + uint64_t fill; + __m64 vfill; + uint32_t byte_width; + uint8_t *byte_line; + +#ifdef __GNUC__ + __m64 v1, v2, v3, v4, v5, v6, v7; +#endif + + if (bpp != 16 && bpp != 32 && bpp != 8) + return FALSE; + + if (bpp == 8) + { + stride = stride * (int) sizeof (uint32_t) / 1; + byte_line = (uint8_t *)(((uint8_t *)bits) + stride * y + x); + byte_width = width; + stride *= 1; + xor = (xor & 0xff) * 0x01010101; + } + else if (bpp == 16) + { + stride = stride * (int) sizeof (uint32_t) / 2; + byte_line = (uint8_t *)(((uint16_t *)bits) + stride * y + x); + byte_width = 2 * width; + stride *= 2; + xor = (xor & 0xffff) * 0x00010001; + } + else + { + stride = stride * (int) sizeof (uint32_t) / 4; + byte_line = (uint8_t *)(((uint32_t *)bits) + stride * y + x); + byte_width = 4 * width; + stride *= 4; + } + + fill = ((uint64_t)xor << 32) | xor; + vfill = to_m64 (fill); + +#ifdef __GNUC__ + __asm__ ( + "movq %7, %0\n" + "movq %7, %1\n" + "movq %7, %2\n" + "movq %7, %3\n" + "movq %7, %4\n" + "movq %7, %5\n" + "movq %7, %6\n" + : "=&y" (v1), "=&y" (v2), "=&y" (v3), + "=&y" (v4), "=&y" (v5), "=&y" (v6), "=y" (v7) + : "y" (vfill)); +#endif + + while (height--) + { + int w; + uint8_t *d = byte_line; + + byte_line += stride; + w = byte_width; + + while (w >= 1 && ((unsigned long)d & 1)) + { + *(uint8_t *)d = (xor & 0xff); + w--; + d++; + } + + while (w >= 2 && ((unsigned long)d & 3)) + { + *(uint16_t *)d = xor; + w -= 2; + d += 2; + } + + while (w >= 4 && ((unsigned long)d & 7)) + { + *(uint32_t *)d = xor; + + w -= 4; + d += 4; + } + + while (w >= 64) + { +#ifdef __GNUC__ + __asm__ ( + "movq %1, (%0)\n" + "movq %2, 8(%0)\n" + "movq %3, 16(%0)\n" + "movq %4, 24(%0)\n" + "movq %5, 32(%0)\n" + "movq %6, 40(%0)\n" + "movq %7, 48(%0)\n" + "movq %8, 56(%0)\n" + : + : "r" (d), + "y" (vfill), "y" (v1), "y" (v2), "y" (v3), + "y" (v4), "y" (v5), "y" (v6), "y" (v7) + : "memory"); +#else + *(__m64*) (d + 0) = vfill; + *(__m64*) (d + 8) = vfill; + *(__m64*) (d + 16) = vfill; + *(__m64*) (d + 24) = vfill; + *(__m64*) (d + 32) = vfill; + *(__m64*) (d + 40) = vfill; + *(__m64*) (d + 48) = vfill; + *(__m64*) (d + 56) = vfill; +#endif + w -= 64; + d += 64; + } + + while (w >= 4) + { + *(uint32_t *)d = xor; + + w -= 4; + d += 4; + } + while (w >= 2) + { + *(uint16_t *)d = xor; + w -= 2; + d += 2; + } + while (w >= 1) + { + *(uint8_t *)d = (xor & 0xff); + w--; + d++; + } + + } + + _mm_empty (); + return TRUE; +} + +static void +mmx_composite_src_n_8_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t src, srca; + uint32_t *dst_line, *dst; + uint8_t *mask_line, *mask; + int dst_stride, mask_stride; + int32_t w; + __m64 vsrc, vsrca; + uint64_t srcsrc; + + CHECKPOINT (); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + srca = src >> 24; + if (src == 0) + { + pixman_fill_mmx (dst_image->bits.bits, dst_image->bits.rowstride, + PIXMAN_FORMAT_BPP (dst_image->bits.format), + dest_x, dest_y, width, height, 0); + return; + } + + srcsrc = (uint64_t)src << 32 | src; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + mask = mask_line; + mask_line += mask_stride; + w = width; + + CHECKPOINT (); + + while (w && (unsigned long)dst & 7) + { + uint64_t m = *mask; + + if (m) + { + __m64 vdest = in (vsrc, expand_alpha_rev (to_m64 (m))); + + *dst = store8888 (vdest); + } + else + { + *dst = 0; + } + + w--; + mask++; + dst++; + } + + CHECKPOINT (); + + while (w >= 2) + { + uint64_t m0, m1; + m0 = *mask; + m1 = *(mask + 1); + + if (srca == 0xff && (m0 & m1) == 0xff) + { + *(uint64_t *)dst = srcsrc; + } + else if (m0 | m1) + { + __m64 vdest; + __m64 dest0, dest1; + + vdest = *(__m64 *)dst; + + dest0 = in (vsrc, expand_alpha_rev (to_m64 (m0))); + dest1 = in (vsrc, expand_alpha_rev (to_m64 (m1))); + + *(__m64 *)dst = pack8888 (dest0, dest1); + } + else + { + *(uint64_t *)dst = 0; + } + + mask += 2; + dst += 2; + w -= 2; + } + + CHECKPOINT (); + + while (w) + { + uint64_t m = *mask; + + if (m) + { + __m64 vdest = load8888 (*dst); + + vdest = in (vsrc, expand_alpha_rev (to_m64 (m))); + *dst = store8888 (vdest); + } + else + { + *dst = 0; + } + + w--; + mask++; + dst++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_n_8_0565 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t src, srca; + uint16_t *dst_line, *dst; + uint8_t *mask_line, *mask; + int dst_stride, mask_stride; + int32_t w; + __m64 vsrc, vsrca, tmp; + uint64_t srcsrcsrcsrc, src16; + + CHECKPOINT (); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + srca = src >> 24; + if (src == 0) + return; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + tmp = pack_565 (vsrc, _mm_setzero_si64 (), 0); + src16 = to_uint64 (tmp); + + srcsrcsrcsrc = + (uint64_t)src16 << 48 | (uint64_t)src16 << 32 | + (uint64_t)src16 << 16 | (uint64_t)src16; + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + mask = mask_line; + mask_line += mask_stride; + w = width; + + CHECKPOINT (); + + while (w && (unsigned long)dst & 7) + { + uint64_t m = *mask; + + if (m) + { + uint64_t d = *dst; + __m64 vd = to_m64 (d); + __m64 vdest = in_over ( + vsrc, vsrca, expand_alpha_rev (to_m64 (m)), expand565 (vd, 0)); + + vd = pack_565 (vdest, _mm_setzero_si64 (), 0); + *dst = to_uint64 (vd); + } + + w--; + mask++; + dst++; + } + + CHECKPOINT (); + + while (w >= 4) + { + uint64_t m0, m1, m2, m3; + m0 = *mask; + m1 = *(mask + 1); + m2 = *(mask + 2); + m3 = *(mask + 3); + + if (srca == 0xff && (m0 & m1 & m2 & m3) == 0xff) + { + *(uint64_t *)dst = srcsrcsrcsrc; + } + else if (m0 | m1 | m2 | m3) + { + __m64 vdest; + __m64 vm0, vm1, vm2, vm3; + + vdest = *(__m64 *)dst; + + vm0 = to_m64 (m0); + vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm0), + expand565 (vdest, 0)), vdest, 0); + vm1 = to_m64 (m1); + vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm1), + expand565 (vdest, 1)), vdest, 1); + vm2 = to_m64 (m2); + vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm2), + expand565 (vdest, 2)), vdest, 2); + vm3 = to_m64 (m3); + vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm3), + expand565 (vdest, 3)), vdest, 3); + + *(__m64 *)dst = vdest; + } + + w -= 4; + mask += 4; + dst += 4; + } + + CHECKPOINT (); + + while (w) + { + uint64_t m = *mask; + + if (m) + { + uint64_t d = *dst; + __m64 vd = to_m64 (d); + __m64 vdest = in_over (vsrc, vsrca, expand_alpha_rev (to_m64 (m)), + expand565 (vd, 0)); + vd = pack_565 (vdest, _mm_setzero_si64 (), 0); + *dst = to_uint64 (vd); + } + + w--; + mask++; + dst++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_pixbuf_0565 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint16_t *dst_line, *dst; + uint32_t *src_line, *src; + int dst_stride, src_stride; + int32_t w; + + CHECKPOINT (); + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); + +#if 0 + /* FIXME */ + assert (src_image->drawable == mask_image->drawable); +#endif + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + CHECKPOINT (); + + while (w && (unsigned long)dst & 7) + { + __m64 vsrc = load8888 (*src); + uint64_t d = *dst; + __m64 vdest = expand565 (to_m64 (d), 0); + + vdest = pack_565 (over_rev_non_pre (vsrc, vdest), vdest, 0); + + *dst = to_uint64 (vdest); + + w--; + dst++; + src++; + } + + CHECKPOINT (); + + while (w >= 4) + { + uint32_t s0, s1, s2, s3; + unsigned char a0, a1, a2, a3; + + s0 = *src; + s1 = *(src + 1); + s2 = *(src + 2); + s3 = *(src + 3); + + a0 = (s0 >> 24); + a1 = (s1 >> 24); + a2 = (s2 >> 24); + a3 = (s3 >> 24); + + if ((a0 & a1 & a2 & a3) == 0xFF) + { + __m64 vdest; + vdest = pack_565 (invert_colors (load8888 (s0)), _mm_setzero_si64 (), 0); + vdest = pack_565 (invert_colors (load8888 (s1)), vdest, 1); + vdest = pack_565 (invert_colors (load8888 (s2)), vdest, 2); + vdest = pack_565 (invert_colors (load8888 (s3)), vdest, 3); + + *(__m64 *)dst = vdest; + } + else if (s0 | s1 | s2 | s3) + { + __m64 vdest = *(__m64 *)dst; + + vdest = pack_565 (over_rev_non_pre (load8888 (s0), expand565 (vdest, 0)), vdest, 0); + vdest = pack_565 (over_rev_non_pre (load8888 (s1), expand565 (vdest, 1)), vdest, 1); + vdest = pack_565 (over_rev_non_pre (load8888 (s2), expand565 (vdest, 2)), vdest, 2); + vdest = pack_565 (over_rev_non_pre (load8888 (s3), expand565 (vdest, 3)), vdest, 3); + + *(__m64 *)dst = vdest; + } + + w -= 4; + dst += 4; + src += 4; + } + + CHECKPOINT (); + + while (w) + { + __m64 vsrc = load8888 (*src); + uint64_t d = *dst; + __m64 vdest = expand565 (to_m64 (d), 0); + + vdest = pack_565 (over_rev_non_pre (vsrc, vdest), vdest, 0); + + *dst = to_uint64 (vdest); + + w--; + dst++; + src++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_pixbuf_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t *dst_line, *dst; + uint32_t *src_line, *src; + int dst_stride, src_stride; + int32_t w; + + CHECKPOINT (); + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); + +#if 0 + /* FIXME */ + assert (src_image->drawable == mask_image->drawable); +#endif + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + while (w && (unsigned long)dst & 7) + { + __m64 s = load8888 (*src); + __m64 d = load8888 (*dst); + + *dst = store8888 (over_rev_non_pre (s, d)); + + w--; + dst++; + src++; + } + + while (w >= 2) + { + uint64_t s0, s1; + unsigned char a0, a1; + __m64 d0, d1; + + s0 = *src; + s1 = *(src + 1); + + a0 = (s0 >> 24); + a1 = (s1 >> 24); + + if ((a0 & a1) == 0xFF) + { + d0 = invert_colors (load8888 (s0)); + d1 = invert_colors (load8888 (s1)); + + *(__m64 *)dst = pack8888 (d0, d1); + } + else if (s0 | s1) + { + __m64 vdest = *(__m64 *)dst; + + d0 = over_rev_non_pre (load8888 (s0), expand8888 (vdest, 0)); + d1 = over_rev_non_pre (load8888 (s1), expand8888 (vdest, 1)); + + *(__m64 *)dst = pack8888 (d0, d1); + } + + w -= 2; + dst += 2; + src += 2; + } + + while (w) + { + __m64 s = load8888 (*src); + __m64 d = load8888 (*dst); + + *dst = store8888 (over_rev_non_pre (s, d)); + + w--; + dst++; + src++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_over_n_8888_0565_ca (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t src, srca; + uint16_t *dst_line; + uint32_t *mask_line; + int dst_stride, mask_stride; + __m64 vsrc, vsrca; + + CHECKPOINT (); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + srca = src >> 24; + if (src == 0) + return; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint32_t, mask_stride, mask_line, 1); + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + while (height--) + { + int twidth = width; + uint32_t *p = (uint32_t *)mask_line; + uint16_t *q = (uint16_t *)dst_line; + + while (twidth && ((unsigned long)q & 7)) + { + uint32_t m = *(uint32_t *)p; + + if (m) + { + uint64_t d = *q; + __m64 vdest = expand565 (to_m64 (d), 0); + vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m), vdest), vdest, 0); + *q = to_uint64 (vdest); + } + + twidth--; + p++; + q++; + } + + while (twidth >= 4) + { + uint32_t m0, m1, m2, m3; + + m0 = *p; + m1 = *(p + 1); + m2 = *(p + 2); + m3 = *(p + 3); + + if ((m0 | m1 | m2 | m3)) + { + __m64 vdest = *(__m64 *)q; + + vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m0), expand565 (vdest, 0)), vdest, 0); + vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m1), expand565 (vdest, 1)), vdest, 1); + vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m2), expand565 (vdest, 2)), vdest, 2); + vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m3), expand565 (vdest, 3)), vdest, 3); + + *(__m64 *)q = vdest; + } + twidth -= 4; + p += 4; + q += 4; + } + + while (twidth) + { + uint32_t m; + + m = *(uint32_t *)p; + if (m) + { + uint64_t d = *q; + __m64 vdest = expand565 (to_m64 (d), 0); + vdest = pack_565 (in_over (vsrc, vsrca, load8888 (m), vdest), vdest, 0); + *q = to_uint64 (vdest); + } + + twidth--; + p++; + q++; + } + + mask_line += mask_stride; + dst_line += dst_stride; + } + + _mm_empty (); +} + +static void +mmx_composite_in_n_8_8 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint8_t *dst_line, *dst; + uint8_t *mask_line, *mask; + int dst_stride, mask_stride; + int32_t w; + uint32_t src; + uint8_t sa; + __m64 vsrc, vsrca; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + sa = src >> 24; + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + mask = mask_line; + mask_line += mask_stride; + w = width; + + if ((((unsigned long)dst_image & 3) == 0) && + (((unsigned long)src_image & 3) == 0)) + { + while (w >= 4) + { + uint32_t m; + __m64 vmask; + __m64 vdest; + + m = 0; + + vmask = load8888 (*(uint32_t *)mask); + vdest = load8888 (*(uint32_t *)dst); + + *(uint32_t *)dst = store8888 (in (in (vsrca, vmask), vdest)); + + dst += 4; + mask += 4; + w -= 4; + } + } + + while (w--) + { + uint16_t tmp; + uint8_t a; + uint32_t m, d; + + a = *mask++; + d = *dst; + + m = MUL_UN8 (sa, a, tmp); + d = MUL_UN8 (m, d, tmp); + + *dst++ = d; + } + } + + _mm_empty (); +} + +static void +mmx_composite_in_8_8 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint8_t *dst_line, *dst; + uint8_t *src_line, *src; + int src_stride, dst_stride; + int32_t w; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint8_t, src_stride, src_line, 1); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + if ((((unsigned long)dst_image & 3) == 0) && + (((unsigned long)src_image & 3) == 0)) + { + while (w >= 4) + { + uint32_t *s = (uint32_t *)src; + uint32_t *d = (uint32_t *)dst; + + *d = store8888 (in (load8888 (*s), load8888 (*d))); + + w -= 4; + dst += 4; + src += 4; + } + } + + while (w--) + { + uint8_t s, d; + uint16_t tmp; + + s = *src; + d = *dst; + + *dst = MUL_UN8 (s, d, tmp); + + src++; + dst++; + } + } + + _mm_empty (); +} + +static void +mmx_composite_add_n_8_8 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint8_t *dst_line, *dst; + uint8_t *mask_line, *mask; + int dst_stride, mask_stride; + int32_t w; + uint32_t src; + uint8_t sa; + __m64 vsrc, vsrca; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); + + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); + + sa = src >> 24; + + if (src == 0) + return; + + vsrc = load8888 (src); + vsrca = expand_alpha (vsrc); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + mask = mask_line; + mask_line += mask_stride; + w = width; + + if ((((unsigned long)mask_image & 3) == 0) && + (((unsigned long)dst_image & 3) == 0)) + { + while (w >= 4) + { + __m64 vmask = load8888 (*(uint32_t *)mask); + __m64 vdest = load8888 (*(uint32_t *)dst); + + *(uint32_t *)dst = store8888 (_mm_adds_pu8 (in (vsrca, vmask), vdest)); + + w -= 4; + dst += 4; + mask += 4; + } + } + + while (w--) + { + uint16_t tmp; + uint16_t a; + uint32_t m, d; + uint32_t r; + + a = *mask++; + d = *dst; + + m = MUL_UN8 (sa, a, tmp); + r = ADD_UN8 (m, d, tmp); + + *dst++ = r; + } + } + + _mm_empty (); +} + +static void +mmx_composite_add_8_8 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint8_t *dst_line, *dst; + uint8_t *src_line, *src; + int dst_stride, src_stride; + int32_t w; + uint8_t s, d; + uint16_t t; + + CHECKPOINT (); + + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint8_t, src_stride, src_line, 1); + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + while (w && (unsigned long)dst & 7) + { + s = *src; + d = *dst; + t = d + s; + s = t | (0 - (t >> 8)); + *dst = s; + + dst++; + src++; + w--; + } + + while (w >= 8) + { + *(__m64*)dst = _mm_adds_pu8 (*(__m64*)src, *(__m64*)dst); + dst += 8; + src += 8; + w -= 8; + } + + while (w) + { + s = *src; + d = *dst; + t = d + s; + s = t | (0 - (t >> 8)); + *dst = s; + + dst++; + src++; + w--; + } + } + + _mm_empty (); +} + +static void +mmx_composite_add_8888_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + __m64 dst64; + uint32_t *dst_line, *dst; + uint32_t *src_line, *src; + int dst_stride, src_stride; + int32_t w; + + CHECKPOINT (); + + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + + while (height--) + { + dst = dst_line; + dst_line += dst_stride; + src = src_line; + src_line += src_stride; + w = width; + + while (w && (unsigned long)dst & 7) + { + *dst = _mm_cvtsi64_si32 (_mm_adds_pu8 (_mm_cvtsi32_si64 (*src), + _mm_cvtsi32_si64 (*dst))); + dst++; + src++; + w--; + } + + while (w >= 2) + { + dst64 = _mm_adds_pu8 (*(__m64*)src, *(__m64*)dst); + *(uint64_t*)dst = to_uint64 (dst64); + dst += 2; + src += 2; + w -= 2; + } + + if (w) + { + *dst = _mm_cvtsi64_si32 (_mm_adds_pu8 (_mm_cvtsi32_si64 (*src), + _mm_cvtsi32_si64 (*dst))); + + } + } + + _mm_empty (); +} + +static pixman_bool_t +pixman_blt_mmx (uint32_t *src_bits, + uint32_t *dst_bits, + int src_stride, + int dst_stride, + int src_bpp, + int dst_bpp, + int src_x, + int src_y, + int dst_x, + int dst_y, + int width, + int height) +{ + uint8_t * src_bytes; + uint8_t * dst_bytes; + int byte_width; + + if (src_bpp != dst_bpp) + return FALSE; + + if (src_bpp == 16) + { + src_stride = src_stride * (int) sizeof (uint32_t) / 2; + dst_stride = dst_stride * (int) sizeof (uint32_t) / 2; + src_bytes = (uint8_t *)(((uint16_t *)src_bits) + src_stride * (src_y) + (src_x)); + dst_bytes = (uint8_t *)(((uint16_t *)dst_bits) + dst_stride * (dst_y) + (dst_x)); + byte_width = 2 * width; + src_stride *= 2; + dst_stride *= 2; + } + else if (src_bpp == 32) + { + src_stride = src_stride * (int) sizeof (uint32_t) / 4; + dst_stride = dst_stride * (int) sizeof (uint32_t) / 4; + src_bytes = (uint8_t *)(((uint32_t *)src_bits) + src_stride * (src_y) + (src_x)); + dst_bytes = (uint8_t *)(((uint32_t *)dst_bits) + dst_stride * (dst_y) + (dst_x)); + byte_width = 4 * width; + src_stride *= 4; + dst_stride *= 4; + } + else + { + return FALSE; + } + + while (height--) + { + int w; + uint8_t *s = src_bytes; + uint8_t *d = dst_bytes; + src_bytes += src_stride; + dst_bytes += dst_stride; + w = byte_width; + + while (w >= 2 && ((unsigned long)d & 3)) + { + *(uint16_t *)d = *(uint16_t *)s; + w -= 2; + s += 2; + d += 2; + } + + while (w >= 4 && ((unsigned long)d & 7)) + { + *(uint32_t *)d = *(uint32_t *)s; + + w -= 4; + s += 4; + d += 4; + } + + while (w >= 64) + { +#if defined (__GNUC__) || (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x590)) + __asm__ ( + "movq (%1), %%mm0\n" + "movq 8(%1), %%mm1\n" + "movq 16(%1), %%mm2\n" + "movq 24(%1), %%mm3\n" + "movq 32(%1), %%mm4\n" + "movq 40(%1), %%mm5\n" + "movq 48(%1), %%mm6\n" + "movq 56(%1), %%mm7\n" + + "movq %%mm0, (%0)\n" + "movq %%mm1, 8(%0)\n" + "movq %%mm2, 16(%0)\n" + "movq %%mm3, 24(%0)\n" + "movq %%mm4, 32(%0)\n" + "movq %%mm5, 40(%0)\n" + "movq %%mm6, 48(%0)\n" + "movq %%mm7, 56(%0)\n" + : + : "r" (d), "r" (s) + : "memory", + "%mm0", "%mm1", "%mm2", "%mm3", + "%mm4", "%mm5", "%mm6", "%mm7"); +#else + __m64 v0 = *(__m64 *)(s + 0); + __m64 v1 = *(__m64 *)(s + 8); + __m64 v2 = *(__m64 *)(s + 16); + __m64 v3 = *(__m64 *)(s + 24); + __m64 v4 = *(__m64 *)(s + 32); + __m64 v5 = *(__m64 *)(s + 40); + __m64 v6 = *(__m64 *)(s + 48); + __m64 v7 = *(__m64 *)(s + 56); + *(__m64 *)(d + 0) = v0; + *(__m64 *)(d + 8) = v1; + *(__m64 *)(d + 16) = v2; + *(__m64 *)(d + 24) = v3; + *(__m64 *)(d + 32) = v4; + *(__m64 *)(d + 40) = v5; + *(__m64 *)(d + 48) = v6; + *(__m64 *)(d + 56) = v7; +#endif + + w -= 64; + s += 64; + d += 64; + } + while (w >= 4) + { + *(uint32_t *)d = *(uint32_t *)s; + + w -= 4; + s += 4; + d += 4; + } + if (w >= 2) + { + *(uint16_t *)d = *(uint16_t *)s; + w -= 2; + s += 2; + d += 2; + } + } + + _mm_empty (); + + return TRUE; +} + +static void +mmx_composite_copy_area (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + pixman_blt_mmx (src_image->bits.bits, + dst_image->bits.bits, + src_image->bits.rowstride, + dst_image->bits.rowstride, + PIXMAN_FORMAT_BPP (src_image->bits.format), + PIXMAN_FORMAT_BPP (dst_image->bits.format), + src_x, src_y, dest_x, dest_y, width, height); +} + +#if 0 +static void +mmx_composite_over_x888_8_8888 (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src_image, + pixman_image_t * mask_image, + pixman_image_t * dst_image, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height) +{ + uint32_t *src, *src_line; + uint32_t *dst, *dst_line; + uint8_t *mask, *mask_line; + int src_stride, mask_stride, dst_stride; + int32_t w; + + PIXMAN_IMAGE_GET_LINE (dst_image, dest_x, dest_y, uint32_t, dst_stride, dst_line, 1); + PIXMAN_IMAGE_GET_LINE (mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); + PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); + + while (height--) + { + src = src_line; + src_line += src_stride; + dst = dst_line; + dst_line += dst_stride; + mask = mask_line; + mask_line += mask_stride; + + w = width; + + while (w--) + { + uint64_t m = *mask; + + if (m) + { + __m64 s = load8888 (*src | 0xff000000); + + if (m == 0xff) + { + *dst = store8888 (s); + } + else + { + __m64 sa = expand_alpha (s); + __m64 vm = expand_alpha_rev (to_m64 (m)); + __m64 vdest = in_over (s, sa, vm, load8888 (*dst)); + + *dst = store8888 (vdest); + } + } + + mask++; + dst++; + src++; + } + } + + _mm_empty (); +} +#endif + +static const pixman_fast_path_t mmx_fast_paths[] = +{ + PIXMAN_STD_FAST_PATH (OVER, solid, a8, r5g6b5, mmx_composite_over_n_8_0565 ), + PIXMAN_STD_FAST_PATH (OVER, solid, a8, b5g6r5, mmx_composite_over_n_8_0565 ), + PIXMAN_STD_FAST_PATH (OVER, solid, a8, a8r8g8b8, mmx_composite_over_n_8_8888 ), + PIXMAN_STD_FAST_PATH (OVER, solid, a8, x8r8g8b8, mmx_composite_over_n_8_8888 ), + PIXMAN_STD_FAST_PATH (OVER, solid, a8, a8b8g8r8, mmx_composite_over_n_8_8888 ), + PIXMAN_STD_FAST_PATH (OVER, solid, a8, x8b8g8r8, mmx_composite_over_n_8_8888 ), + PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8r8g8b8, a8r8g8b8, mmx_composite_over_n_8888_8888_ca ), + PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8r8g8b8, x8r8g8b8, mmx_composite_over_n_8888_8888_ca ), + PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8r8g8b8, r5g6b5, mmx_composite_over_n_8888_0565_ca ), + PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8b8g8r8, a8b8g8r8, mmx_composite_over_n_8888_8888_ca ), + PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8b8g8r8, x8b8g8r8, mmx_composite_over_n_8888_8888_ca ), + PIXMAN_STD_FAST_PATH_CA (OVER, solid, a8b8g8r8, b5g6r5, mmx_composite_over_n_8888_0565_ca ), + PIXMAN_STD_FAST_PATH (OVER, pixbuf, pixbuf, a8r8g8b8, mmx_composite_over_pixbuf_8888 ), + PIXMAN_STD_FAST_PATH (OVER, pixbuf, pixbuf, x8r8g8b8, mmx_composite_over_pixbuf_8888 ), + PIXMAN_STD_FAST_PATH (OVER, pixbuf, pixbuf, r5g6b5, mmx_composite_over_pixbuf_0565 ), + PIXMAN_STD_FAST_PATH (OVER, rpixbuf, rpixbuf, a8b8g8r8, mmx_composite_over_pixbuf_8888 ), + PIXMAN_STD_FAST_PATH (OVER, rpixbuf, rpixbuf, x8b8g8r8, mmx_composite_over_pixbuf_8888 ), + PIXMAN_STD_FAST_PATH (OVER, rpixbuf, rpixbuf, b5g6r5, mmx_composite_over_pixbuf_0565 ), + PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, solid, a8r8g8b8, mmx_composite_over_x888_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, solid, x8r8g8b8, mmx_composite_over_x888_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, x8b8g8r8, solid, a8b8g8r8, mmx_composite_over_x888_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, x8b8g8r8, solid, x8b8g8r8, mmx_composite_over_x888_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, solid, a8r8g8b8, mmx_composite_over_8888_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, solid, x8r8g8b8, mmx_composite_over_8888_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, solid, a8b8g8r8, mmx_composite_over_8888_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, solid, x8b8g8r8, mmx_composite_over_8888_n_8888 ), +#if 0 + /* FIXME: This code is commented out since it's apparently + * not actually faster than the generic code. + */ + PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, a8, x8r8g8b8, mmx_composite_over_x888_8_8888 ), + PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, a8, a8r8g8b8, mmx_composite_over_x888_8_8888 ), + PIXMAN_STD_FAST_PATH (OVER, x8b8r8g8, a8, x8b8g8r8, mmx_composite_over_x888_8_8888 ), + PIXMAN_STD_FAST_PATH (OVER, x8b8r8g8, a8, a8r8g8b8, mmx_composite_over_x888_8_8888 ), +#endif + PIXMAN_STD_FAST_PATH (OVER, solid, null, a8r8g8b8, mmx_composite_over_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, solid, null, x8r8g8b8, mmx_composite_over_n_8888 ), + PIXMAN_STD_FAST_PATH (OVER, solid, null, r5g6b5, mmx_composite_over_n_0565 ), + PIXMAN_STD_FAST_PATH (OVER, x8r8g8b8, null, x8r8g8b8, mmx_composite_copy_area ), + PIXMAN_STD_FAST_PATH (OVER, x8b8g8r8, null, x8b8g8r8, mmx_composite_copy_area ), + + PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, null, a8r8g8b8, mmx_composite_over_8888_8888 ), + PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, null, x8r8g8b8, mmx_composite_over_8888_8888 ), + PIXMAN_STD_FAST_PATH (OVER, a8r8g8b8, null, r5g6b5, mmx_composite_over_8888_0565 ), + PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, null, a8b8g8r8, mmx_composite_over_8888_8888 ), + PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, null, x8b8g8r8, mmx_composite_over_8888_8888 ), + PIXMAN_STD_FAST_PATH (OVER, a8b8g8r8, null, b5g6r5, mmx_composite_over_8888_0565 ), + + PIXMAN_STD_FAST_PATH (ADD, a8r8g8b8, null, a8r8g8b8, mmx_composite_add_8888_8888 ), + PIXMAN_STD_FAST_PATH (ADD, a8b8g8r8, null, a8b8g8r8, mmx_composite_add_8888_8888 ), + PIXMAN_STD_FAST_PATH (ADD, a8, null, a8, mmx_composite_add_8_8 ), + PIXMAN_STD_FAST_PATH (ADD, solid, a8, a8, mmx_composite_add_n_8_8 ), + + PIXMAN_STD_FAST_PATH (SRC, solid, a8, a8r8g8b8, mmx_composite_src_n_8_8888 ), + PIXMAN_STD_FAST_PATH (SRC, solid, a8, x8r8g8b8, mmx_composite_src_n_8_8888 ), + PIXMAN_STD_FAST_PATH (SRC, solid, a8, a8b8g8r8, mmx_composite_src_n_8_8888 ), + PIXMAN_STD_FAST_PATH (SRC, solid, a8, x8b8g8r8, mmx_composite_src_n_8_8888 ), + PIXMAN_STD_FAST_PATH (SRC, a8r8g8b8, null, a8r8g8b8, mmx_composite_copy_area ), + PIXMAN_STD_FAST_PATH (SRC, a8b8g8r8, null, a8b8g8r8, mmx_composite_copy_area ), + PIXMAN_STD_FAST_PATH (SRC, a8r8g8b8, null, x8r8g8b8, mmx_composite_copy_area ), + PIXMAN_STD_FAST_PATH (SRC, a8b8g8r8, null, x8b8g8r8, mmx_composite_copy_area ), + PIXMAN_STD_FAST_PATH (SRC, x8r8g8b8, null, x8r8g8b8, mmx_composite_copy_area ), + PIXMAN_STD_FAST_PATH (SRC, x8b8g8r8, null, x8b8g8r8, mmx_composite_copy_area ), + PIXMAN_STD_FAST_PATH (SRC, r5g6b5, null, r5g6b5, mmx_composite_copy_area ), + PIXMAN_STD_FAST_PATH (SRC, b5g6r5, null, b5g6r5, mmx_composite_copy_area ), + + PIXMAN_STD_FAST_PATH (IN, a8, null, a8, mmx_composite_in_8_8 ), + PIXMAN_STD_FAST_PATH (IN, solid, a8, a8, mmx_composite_in_n_8_8 ), + + { PIXMAN_OP_NONE }, +}; + +static pixman_bool_t +mmx_blt (pixman_implementation_t *imp, + uint32_t * src_bits, + uint32_t * dst_bits, + int src_stride, + int dst_stride, + int src_bpp, + int dst_bpp, + int src_x, + int src_y, + int dst_x, + int dst_y, + int width, + int height) +{ + if (!pixman_blt_mmx ( + src_bits, dst_bits, src_stride, dst_stride, src_bpp, dst_bpp, + src_x, src_y, dst_x, dst_y, width, height)) + + { + return _pixman_implementation_blt ( + imp->delegate, + src_bits, dst_bits, src_stride, dst_stride, src_bpp, dst_bpp, + src_x, src_y, dst_x, dst_y, width, height); + } + + return TRUE; +} + +static pixman_bool_t +mmx_fill (pixman_implementation_t *imp, + uint32_t * bits, + int stride, + int bpp, + int x, + int y, + int width, + int height, + uint32_t xor) +{ + if (!pixman_fill_mmx (bits, stride, bpp, x, y, width, height, xor)) + { + return _pixman_implementation_fill ( + imp->delegate, bits, stride, bpp, x, y, width, height, xor); + } + + return TRUE; +} + +pixman_implementation_t * +_pixman_implementation_create_mmx (void) +{ + pixman_implementation_t *general = _pixman_implementation_create_fast_path (); + pixman_implementation_t *imp = _pixman_implementation_create (general, mmx_fast_paths); + + imp->combine_32[PIXMAN_OP_OVER] = mmx_combine_over_u; + imp->combine_32[PIXMAN_OP_OVER_REVERSE] = mmx_combine_over_reverse_u; + imp->combine_32[PIXMAN_OP_IN] = mmx_combine_in_u; + imp->combine_32[PIXMAN_OP_IN_REVERSE] = mmx_combine_in_reverse_u; + imp->combine_32[PIXMAN_OP_OUT] = mmx_combine_out_u; + imp->combine_32[PIXMAN_OP_OUT_REVERSE] = mmx_combine_out_reverse_u; + imp->combine_32[PIXMAN_OP_ATOP] = mmx_combine_atop_u; + imp->combine_32[PIXMAN_OP_ATOP_REVERSE] = mmx_combine_atop_reverse_u; + imp->combine_32[PIXMAN_OP_XOR] = mmx_combine_xor_u; + imp->combine_32[PIXMAN_OP_ADD] = mmx_combine_add_u; + imp->combine_32[PIXMAN_OP_SATURATE] = mmx_combine_saturate_u; + + imp->combine_32_ca[PIXMAN_OP_SRC] = mmx_combine_src_ca; + imp->combine_32_ca[PIXMAN_OP_OVER] = mmx_combine_over_ca; + imp->combine_32_ca[PIXMAN_OP_OVER_REVERSE] = mmx_combine_over_reverse_ca; + imp->combine_32_ca[PIXMAN_OP_IN] = mmx_combine_in_ca; + imp->combine_32_ca[PIXMAN_OP_IN_REVERSE] = mmx_combine_in_reverse_ca; + imp->combine_32_ca[PIXMAN_OP_OUT] = mmx_combine_out_ca; + imp->combine_32_ca[PIXMAN_OP_OUT_REVERSE] = mmx_combine_out_reverse_ca; + imp->combine_32_ca[PIXMAN_OP_ATOP] = mmx_combine_atop_ca; + imp->combine_32_ca[PIXMAN_OP_ATOP_REVERSE] = mmx_combine_atop_reverse_ca; + imp->combine_32_ca[PIXMAN_OP_XOR] = mmx_combine_xor_ca; + imp->combine_32_ca[PIXMAN_OP_ADD] = mmx_combine_add_ca; + + imp->blt = mmx_blt; + imp->fill = mmx_fill; + + return imp; +} + +#endif /* USE_MMX */ diff --git a/pixman/pixman/pixman-private.h b/pixman/pixman/pixman-private.h index b6eb24835..1662d2c46 100644 --- a/pixman/pixman/pixman-private.h +++ b/pixman/pixman/pixman-private.h @@ -1,867 +1,912 @@ -#ifndef PACKAGE
-# error config.h must be included before pixman-private.h
-#endif
-
-#ifndef PIXMAN_PRIVATE_H
-#define PIXMAN_PRIVATE_H
-
-#define PIXMAN_DISABLE_DEPRECATED
-#define PIXMAN_USE_INTERNAL_API
-
-#include "pixman.h"
-#include <time.h>
-#include <assert.h>
-#include <stdio.h>
-#include <string.h>
-
-#include "pixman-compiler.h"
-
-/*
- * Images
- */
-typedef struct image_common image_common_t;
-typedef struct solid_fill solid_fill_t;
-typedef struct gradient gradient_t;
-typedef struct linear_gradient linear_gradient_t;
-typedef struct horizontal_gradient horizontal_gradient_t;
-typedef struct vertical_gradient vertical_gradient_t;
-typedef struct conical_gradient conical_gradient_t;
-typedef struct radial_gradient radial_gradient_t;
-typedef struct bits_image bits_image_t;
-typedef struct circle circle_t;
-
-typedef void (*fetch_scanline_t) (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t *buffer,
- const uint32_t *mask);
-
-typedef uint32_t (*fetch_pixel_32_t) (bits_image_t *image,
- int x,
- int y);
-
-typedef uint64_t (*fetch_pixel_64_t) (bits_image_t *image,
- int x,
- int y);
-
-typedef void (*store_scanline_t) (bits_image_t * image,
- int x,
- int y,
- int width,
- const uint32_t *values);
-
-typedef enum
-{
- BITS,
- LINEAR,
- CONICAL,
- RADIAL,
- SOLID
-} image_type_t;
-
-typedef enum
-{
- SOURCE_IMAGE_CLASS_UNKNOWN,
- SOURCE_IMAGE_CLASS_HORIZONTAL,
-} source_image_class_t;
-
-typedef source_image_class_t (*classify_func_t) (pixman_image_t *image,
- int x,
- int y,
- int width,
- int height);
-typedef void (*property_changed_func_t) (pixman_image_t *image);
-
-struct image_common
-{
- image_type_t type;
- int32_t ref_count;
- pixman_region32_t clip_region;
- int32_t alpha_count; /* How many times this image is being used as an alpha map */
- pixman_bool_t have_clip_region; /* FALSE if there is no clip */
- pixman_bool_t client_clip; /* Whether the source clip was
- set by a client */
- pixman_bool_t clip_sources; /* Whether the clip applies when
- * the image is used as a source
- */
- pixman_bool_t dirty;
- pixman_transform_t * transform;
- pixman_repeat_t repeat;
- pixman_filter_t filter;
- pixman_fixed_t * filter_params;
- int n_filter_params;
- bits_image_t * alpha_map;
- int alpha_origin_x;
- int alpha_origin_y;
- pixman_bool_t component_alpha;
- classify_func_t classify;
- property_changed_func_t property_changed;
- fetch_scanline_t get_scanline_32;
- fetch_scanline_t get_scanline_64;
-
- pixman_image_destroy_func_t destroy_func;
- void * destroy_data;
-
- uint32_t flags;
- pixman_format_code_t extended_format_code;
-};
-
-struct solid_fill
-{
- image_common_t common;
- pixman_color_t color;
-
- uint32_t color_32;
- uint64_t color_64;
-};
-
-struct gradient
-{
- image_common_t common;
- int n_stops;
- pixman_gradient_stop_t *stops;
-};
-
-struct linear_gradient
-{
- gradient_t common;
- pixman_point_fixed_t p1;
- pixman_point_fixed_t p2;
-};
-
-struct circle
-{
- pixman_fixed_t x;
- pixman_fixed_t y;
- pixman_fixed_t radius;
-};
-
-struct radial_gradient
-{
- gradient_t common;
-
- circle_t c1;
- circle_t c2;
-
- circle_t delta;
- double a;
- double inva;
- double mindr;
-};
-
-struct conical_gradient
-{
- gradient_t common;
- pixman_point_fixed_t center;
- double angle;
-};
-
-struct bits_image
-{
- image_common_t common;
- pixman_format_code_t format;
- const pixman_indexed_t * indexed;
- int width;
- int height;
- uint32_t * bits;
- uint32_t * free_me;
- int rowstride; /* in number of uint32_t's */
-
- fetch_scanline_t fetch_scanline_32;
- fetch_pixel_32_t fetch_pixel_32;
- store_scanline_t store_scanline_32;
-
- fetch_scanline_t fetch_scanline_64;
- fetch_pixel_64_t fetch_pixel_64;
- store_scanline_t store_scanline_64;
-
- /* Used for indirect access to the bits */
- pixman_read_memory_func_t read_func;
- pixman_write_memory_func_t write_func;
-};
-
-union pixman_image
-{
- image_type_t type;
- image_common_t common;
- bits_image_t bits;
- gradient_t gradient;
- linear_gradient_t linear;
- conical_gradient_t conical;
- radial_gradient_t radial;
- solid_fill_t solid;
-};
-
-void
-_pixman_bits_image_setup_accessors (bits_image_t *image);
-
-void
-_pixman_image_get_scanline_generic_64 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *mask);
-
-source_image_class_t
-_pixman_image_classify (pixman_image_t *image,
- int x,
- int y,
- int width,
- int height);
-
-void
-_pixman_image_get_scanline_32 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *mask);
-
-/* Even thought the type of buffer is uint32_t *, the function actually expects
- * a uint64_t *buffer.
- */
-void
-_pixman_image_get_scanline_64 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *unused);
-
-void
-_pixman_image_store_scanline_32 (bits_image_t * image,
- int x,
- int y,
- int width,
- const uint32_t *buffer);
-
-/* Even though the type of buffer is uint32_t *, the function
- * actually expects a uint64_t *buffer.
- */
-void
-_pixman_image_store_scanline_64 (bits_image_t * image,
- int x,
- int y,
- int width,
- const uint32_t *buffer);
-
-pixman_image_t *
-_pixman_image_allocate (void);
-
-pixman_bool_t
-_pixman_init_gradient (gradient_t * gradient,
- const pixman_gradient_stop_t *stops,
- int n_stops);
-void
-_pixman_image_reset_clip_region (pixman_image_t *image);
-
-void
-_pixman_image_validate (pixman_image_t *image);
-
-uint32_t
-_pixman_image_get_solid (pixman_image_t * image,
- pixman_format_code_t format);
-
-#define PIXMAN_IMAGE_GET_LINE(image, x, y, type, out_stride, line, mul) \
- do \
- { \
- uint32_t *__bits__; \
- int __stride__; \
- \
- __bits__ = image->bits.bits; \
- __stride__ = image->bits.rowstride; \
- (out_stride) = \
- __stride__ * (int) sizeof (uint32_t) / (int) sizeof (type); \
- (line) = \
- ((type *) __bits__) + (out_stride) * (y) + (mul) * (x); \
- } while (0)
-
-/*
- * Gradient walker
- */
-typedef struct
-{
- uint32_t left_ag;
- uint32_t left_rb;
- uint32_t right_ag;
- uint32_t right_rb;
- int32_t left_x;
- int32_t right_x;
- int32_t stepper;
-
- pixman_gradient_stop_t *stops;
- int num_stops;
- unsigned int spread;
-
- int need_reset;
-} pixman_gradient_walker_t;
-
-void
-_pixman_gradient_walker_init (pixman_gradient_walker_t *walker,
- gradient_t * gradient,
- unsigned int spread);
-
-void
-_pixman_gradient_walker_reset (pixman_gradient_walker_t *walker,
- pixman_fixed_32_32_t pos);
-
-uint32_t
-_pixman_gradient_walker_pixel (pixman_gradient_walker_t *walker,
- pixman_fixed_32_32_t x);
-
-/*
- * Edges
- */
-
-#define MAX_ALPHA(n) ((1 << (n)) - 1)
-#define N_Y_FRAC(n) ((n) == 1 ? 1 : (1 << ((n) / 2)) - 1)
-#define N_X_FRAC(n) ((n) == 1 ? 1 : (1 << ((n) / 2)) + 1)
-
-#define STEP_Y_SMALL(n) (pixman_fixed_1 / N_Y_FRAC (n))
-#define STEP_Y_BIG(n) (pixman_fixed_1 - (N_Y_FRAC (n) - 1) * STEP_Y_SMALL (n))
-
-#define Y_FRAC_FIRST(n) (STEP_Y_BIG (n) / 2)
-#define Y_FRAC_LAST(n) (Y_FRAC_FIRST (n) + (N_Y_FRAC (n) - 1) * STEP_Y_SMALL (n))
-
-#define STEP_X_SMALL(n) (pixman_fixed_1 / N_X_FRAC (n))
-#define STEP_X_BIG(n) (pixman_fixed_1 - (N_X_FRAC (n) - 1) * STEP_X_SMALL (n))
-
-#define X_FRAC_FIRST(n) (STEP_X_BIG (n) / 2)
-#define X_FRAC_LAST(n) (X_FRAC_FIRST (n) + (N_X_FRAC (n) - 1) * STEP_X_SMALL (n))
-
-#define RENDER_SAMPLES_X(x, n) \
- ((n) == 1? 0 : (pixman_fixed_frac (x) + \
- X_FRAC_FIRST (n)) / STEP_X_SMALL (n))
-
-void
-pixman_rasterize_edges_accessors (pixman_image_t *image,
- pixman_edge_t * l,
- pixman_edge_t * r,
- pixman_fixed_t t,
- pixman_fixed_t b);
-
-/*
- * Implementations
- */
-typedef struct pixman_implementation_t pixman_implementation_t;
-
-typedef void (*pixman_combine_32_func_t) (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width);
-
-typedef void (*pixman_combine_64_func_t) (pixman_implementation_t *imp,
- pixman_op_t op,
- uint64_t * dest,
- const uint64_t * src,
- const uint64_t * mask,
- int width);
-
-typedef void (*pixman_composite_func_t) (pixman_implementation_t *imp,
- pixman_op_t op,
- pixman_image_t * src,
- pixman_image_t * mask,
- pixman_image_t * dest,
- int32_t src_x,
- int32_t src_y,
- int32_t mask_x,
- int32_t mask_y,
- int32_t dest_x,
- int32_t dest_y,
- int32_t width,
- int32_t height);
-typedef pixman_bool_t (*pixman_blt_func_t) (pixman_implementation_t *imp,
- uint32_t * src_bits,
- uint32_t * dst_bits,
- int src_stride,
- int dst_stride,
- int src_bpp,
- int dst_bpp,
- int src_x,
- int src_y,
- int dst_x,
- int dst_y,
- int width,
- int height);
-typedef pixman_bool_t (*pixman_fill_func_t) (pixman_implementation_t *imp,
- uint32_t * bits,
- int stride,
- int bpp,
- int x,
- int y,
- int width,
- int height,
- uint32_t xor);
-
-void _pixman_setup_combiner_functions_32 (pixman_implementation_t *imp);
-void _pixman_setup_combiner_functions_64 (pixman_implementation_t *imp);
-
-typedef struct
-{
- pixman_op_t op;
- pixman_format_code_t src_format;
- uint32_t src_flags;
- pixman_format_code_t mask_format;
- uint32_t mask_flags;
- pixman_format_code_t dest_format;
- uint32_t dest_flags;
- pixman_composite_func_t func;
-} pixman_fast_path_t;
-
-struct pixman_implementation_t
-{
- pixman_implementation_t * toplevel;
- pixman_implementation_t * delegate;
- const pixman_fast_path_t * fast_paths;
-
- pixman_blt_func_t blt;
- pixman_fill_func_t fill;
-
- pixman_combine_32_func_t combine_32[PIXMAN_N_OPERATORS];
- pixman_combine_32_func_t combine_32_ca[PIXMAN_N_OPERATORS];
- pixman_combine_64_func_t combine_64[PIXMAN_N_OPERATORS];
- pixman_combine_64_func_t combine_64_ca[PIXMAN_N_OPERATORS];
-};
-
-pixman_implementation_t *
-_pixman_implementation_create (pixman_implementation_t *delegate,
- const pixman_fast_path_t *fast_paths);
-
-void
-_pixman_implementation_combine_32 (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width);
-void
-_pixman_implementation_combine_64 (pixman_implementation_t *imp,
- pixman_op_t op,
- uint64_t * dest,
- const uint64_t * src,
- const uint64_t * mask,
- int width);
-void
-_pixman_implementation_combine_32_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint32_t * dest,
- const uint32_t * src,
- const uint32_t * mask,
- int width);
-void
-_pixman_implementation_combine_64_ca (pixman_implementation_t *imp,
- pixman_op_t op,
- uint64_t * dest,
- const uint64_t * src,
- const uint64_t * mask,
- int width);
-
-pixman_bool_t
-_pixman_implementation_blt (pixman_implementation_t *imp,
- uint32_t * src_bits,
- uint32_t * dst_bits,
- int src_stride,
- int dst_stride,
- int src_bpp,
- int dst_bpp,
- int src_x,
- int src_y,
- int dst_x,
- int dst_y,
- int width,
- int height);
-
-pixman_bool_t
-_pixman_implementation_fill (pixman_implementation_t *imp,
- uint32_t * bits,
- int stride,
- int bpp,
- int x,
- int y,
- int width,
- int height,
- uint32_t xor);
-
-/* Specific implementations */
-pixman_implementation_t *
-_pixman_implementation_create_general (void);
-
-pixman_implementation_t *
-_pixman_implementation_create_fast_path (void);
-
-#ifdef USE_MMX
-pixman_implementation_t *
-_pixman_implementation_create_mmx (void);
-#endif
-
-#ifdef USE_SSE2
-pixman_implementation_t *
-_pixman_implementation_create_sse2 (void);
-#endif
-
-#ifdef USE_ARM_SIMD
-pixman_implementation_t *
-_pixman_implementation_create_arm_simd (void);
-#endif
-
-#ifdef USE_ARM_NEON
-pixman_implementation_t *
-_pixman_implementation_create_arm_neon (void);
-#endif
-
-#ifdef USE_VMX
-pixman_implementation_t *
-_pixman_implementation_create_vmx (void);
-#endif
-
-pixman_implementation_t *
-_pixman_choose_implementation (void);
-
-
-
-/*
- * Utilities
- */
-
-/* These "formats" all have depth 0, so they
- * will never clash with any real ones
- */
-#define PIXMAN_null PIXMAN_FORMAT (0, 0, 0, 0, 0, 0)
-#define PIXMAN_solid PIXMAN_FORMAT (0, 1, 0, 0, 0, 0)
-#define PIXMAN_pixbuf PIXMAN_FORMAT (0, 2, 0, 0, 0, 0)
-#define PIXMAN_rpixbuf PIXMAN_FORMAT (0, 3, 0, 0, 0, 0)
-#define PIXMAN_unknown PIXMAN_FORMAT (0, 4, 0, 0, 0, 0)
-#define PIXMAN_any PIXMAN_FORMAT (0, 5, 0, 0, 0, 0)
-
-#define PIXMAN_OP_any (PIXMAN_N_OPERATORS + 1)
-
-#define FAST_PATH_ID_TRANSFORM (1 << 0)
-#define FAST_PATH_NO_ALPHA_MAP (1 << 1)
-#define FAST_PATH_NO_CONVOLUTION_FILTER (1 << 2)
-#define FAST_PATH_NO_PAD_REPEAT (1 << 3)
-#define FAST_PATH_NO_REFLECT_REPEAT (1 << 4)
-#define FAST_PATH_NO_ACCESSORS (1 << 5)
-#define FAST_PATH_NARROW_FORMAT (1 << 6)
-#define FAST_PATH_COMPONENT_ALPHA (1 << 8)
-#define FAST_PATH_SAMPLES_OPAQUE (1 << 7)
-#define FAST_PATH_UNIFIED_ALPHA (1 << 9)
-#define FAST_PATH_SCALE_TRANSFORM (1 << 10)
-#define FAST_PATH_NEAREST_FILTER (1 << 11)
-#define FAST_PATH_HAS_TRANSFORM (1 << 12)
-#define FAST_PATH_IS_OPAQUE (1 << 13)
-#define FAST_PATH_NO_NORMAL_REPEAT (1 << 14)
-#define FAST_PATH_NO_NONE_REPEAT (1 << 15)
-#define FAST_PATH_SAMPLES_COVER_CLIP (1 << 16)
-#define FAST_PATH_X_UNIT_POSITIVE (1 << 17)
-#define FAST_PATH_AFFINE_TRANSFORM (1 << 18)
-#define FAST_PATH_Y_UNIT_ZERO (1 << 19)
-#define FAST_PATH_BILINEAR_FILTER (1 << 20)
-
-#define FAST_PATH_PAD_REPEAT \
- (FAST_PATH_NO_NONE_REPEAT | \
- FAST_PATH_NO_NORMAL_REPEAT | \
- FAST_PATH_NO_REFLECT_REPEAT)
-
-#define FAST_PATH_NORMAL_REPEAT \
- (FAST_PATH_NO_NONE_REPEAT | \
- FAST_PATH_NO_PAD_REPEAT | \
- FAST_PATH_NO_REFLECT_REPEAT)
-
-#define FAST_PATH_NONE_REPEAT \
- (FAST_PATH_NO_NORMAL_REPEAT | \
- FAST_PATH_NO_PAD_REPEAT | \
- FAST_PATH_NO_REFLECT_REPEAT)
-
-#define FAST_PATH_REFLECT_REPEAT \
- (FAST_PATH_NO_NONE_REPEAT | \
- FAST_PATH_NO_NORMAL_REPEAT | \
- FAST_PATH_NO_PAD_REPEAT)
-
-#define FAST_PATH_STANDARD_FLAGS \
- (FAST_PATH_NO_CONVOLUTION_FILTER | \
- FAST_PATH_NO_ACCESSORS | \
- FAST_PATH_NO_ALPHA_MAP | \
- FAST_PATH_NARROW_FORMAT)
-
-#define FAST_PATH_STD_DEST_FLAGS \
- (FAST_PATH_NO_ACCESSORS | \
- FAST_PATH_NO_ALPHA_MAP | \
- FAST_PATH_NARROW_FORMAT)
-
-#define SOURCE_FLAGS(format) \
- (FAST_PATH_STANDARD_FLAGS | \
- ((PIXMAN_ ## format == PIXMAN_solid) ? \
- 0 : (FAST_PATH_SAMPLES_COVER_CLIP | FAST_PATH_ID_TRANSFORM)))
-
-#define MASK_FLAGS(format, extra) \
- ((PIXMAN_ ## format == PIXMAN_null) ? 0 : (SOURCE_FLAGS (format) | extra))
-
-#define FAST_PATH(op, src, src_flags, mask, mask_flags, dest, dest_flags, func) \
- PIXMAN_OP_ ## op, \
- PIXMAN_ ## src, \
- src_flags, \
- PIXMAN_ ## mask, \
- mask_flags, \
- PIXMAN_ ## dest, \
- dest_flags, \
- func
-
-#define PIXMAN_STD_FAST_PATH(op, src, mask, dest, func) \
- { FAST_PATH ( \
- op, \
- src, SOURCE_FLAGS (src), \
- mask, MASK_FLAGS (mask, FAST_PATH_UNIFIED_ALPHA), \
- dest, FAST_PATH_STD_DEST_FLAGS, \
- func) }
-
-#define PIXMAN_STD_FAST_PATH_CA(op, src, mask, dest, func) \
- { FAST_PATH ( \
- op, \
- src, SOURCE_FLAGS (src), \
- mask, MASK_FLAGS (mask, FAST_PATH_COMPONENT_ALPHA), \
- dest, FAST_PATH_STD_DEST_FLAGS, \
- func) }
-
-/* Memory allocation helpers */
-void *
-pixman_malloc_ab (unsigned int n, unsigned int b);
-
-void *
-pixman_malloc_abc (unsigned int a, unsigned int b, unsigned int c);
-
-pixman_bool_t
-pixman_multiply_overflows_int (unsigned int a, unsigned int b);
-
-pixman_bool_t
-pixman_addition_overflows_int (unsigned int a, unsigned int b);
-
-/* Compositing utilities */
-void
-pixman_expand (uint64_t * dst,
- const uint32_t * src,
- pixman_format_code_t format,
- int width);
-
-void
-pixman_contract (uint32_t * dst,
- const uint64_t *src,
- int width);
-
-
-/* Region Helpers */
-pixman_bool_t
-pixman_region32_copy_from_region16 (pixman_region32_t *dst,
- pixman_region16_t *src);
-
-pixman_bool_t
-pixman_region16_copy_from_region32 (pixman_region16_t *dst,
- pixman_region32_t *src);
-
-
-/* Misc macros */
-
-#ifndef FALSE
-# define FALSE 0
-#endif
-
-#ifndef TRUE
-# define TRUE 1
-#endif
-
-#ifndef MIN
-# define MIN(a, b) ((a < b) ? a : b)
-#endif
-
-#ifndef MAX
-# define MAX(a, b) ((a > b) ? a : b)
-#endif
-
-/* Integer division that rounds towards -infinity */
-#define DIV(a, b) \
- ((((a) < 0) == ((b) < 0)) ? (a) / (b) : \
- ((a) - (b) + 1 - (((b) < 0) << 1)) / (b))
-
-/* Modulus that produces the remainder wrt. DIV */
-#define MOD(a, b) ((a) < 0 ? ((b) - ((-(a) - 1) % (b))) - 1 : (a) % (b))
-
-#define CLIP(v, low, high) ((v) < (low) ? (low) : ((v) > (high) ? (high) : (v)))
-
-/* Conversion between 8888 and 0565 */
-
-#define CONVERT_8888_TO_0565(s) \
- ((((s) >> 3) & 0x001f) | \
- (((s) >> 5) & 0x07e0) | \
- (((s) >> 8) & 0xf800))
-
-#define CONVERT_0565_TO_0888(s) \
- (((((s) << 3) & 0xf8) | (((s) >> 2) & 0x7)) | \
- ((((s) << 5) & 0xfc00) | (((s) >> 1) & 0x300)) | \
- ((((s) << 8) & 0xf80000) | (((s) << 3) & 0x70000)))
-
-#define CONVERT_0565_TO_8888(s) (CONVERT_0565_TO_0888(s) | 0xff000000)
-
-/* Trivial versions that are useful in macros */
-#define CONVERT_8888_TO_8888(s) (s)
-#define CONVERT_0565_TO_0565(s) (s)
-
-#define PIXMAN_FORMAT_IS_WIDE(f) \
- (PIXMAN_FORMAT_A (f) > 8 || \
- PIXMAN_FORMAT_R (f) > 8 || \
- PIXMAN_FORMAT_G (f) > 8 || \
- PIXMAN_FORMAT_B (f) > 8)
-
-#ifdef WORDS_BIGENDIAN
-# define SCREEN_SHIFT_LEFT(x,n) ((x) << (n))
-# define SCREEN_SHIFT_RIGHT(x,n) ((x) >> (n))
-#else
-# define SCREEN_SHIFT_LEFT(x,n) ((x) >> (n))
-# define SCREEN_SHIFT_RIGHT(x,n) ((x) << (n))
-#endif
-
-/*
- * Various debugging code
- */
-
-#undef DEBUG
-
-#define COMPILE_TIME_ASSERT(x) \
- do { typedef int compile_time_assertion [(x)?1:-1]; } while (0)
-
-/* Turn on debugging depending on what type of release this is
- */
-#if (((PIXMAN_VERSION_MICRO % 2) == 0) && ((PIXMAN_VERSION_MINOR % 2) == 1))
-
-/* Debugging gets turned on for development releases because these
- * are the things that end up in bleeding edge distributions such
- * as Rawhide etc.
- *
- * For performance reasons we don't turn it on for stable releases or
- * random git checkouts. (Random git checkouts are often used for
- * performance work).
- */
-
-# define DEBUG
-
-#endif
-
-#ifdef DEBUG
-
-void
-_pixman_log_error (const char *function, const char *message);
-
-#define return_if_fail(expr) \
- do \
- { \
- if (!(expr)) \
- { \
- _pixman_log_error (FUNC, "The expression " # expr " was false"); \
- return; \
- } \
- } \
- while (0)
-
-#define return_val_if_fail(expr, retval) \
- do \
- { \
- if (!(expr)) \
- { \
- _pixman_log_error (FUNC, "The expression " # expr " was false"); \
- return (retval); \
- } \
- } \
- while (0)
-
-#define critical_if_fail(expr) \
- do \
- { \
- if (!(expr)) \
- _pixman_log_error (FUNC, "The expression " # expr " was false"); \
- } \
- while (0)
-
-
-#else
-
-#define _pixman_log_error(f,m) do { } while (0) \
-
-#define return_if_fail(expr) \
- do \
- { \
- if (!(expr)) \
- return; \
- } \
- while (0)
-
-#define return_val_if_fail(expr, retval) \
- do \
- { \
- if (!(expr)) \
- return (retval); \
- } \
- while (0)
-
-#define critical_if_fail(expr) \
- do \
- { \
- } \
- while (0)
-#endif
-
-/*
- * Timers
- */
-
-#ifdef PIXMAN_TIMERS
-
-static inline uint64_t
-oil_profile_stamp_rdtsc (void)
-{
- uint64_t ts;
-
- __asm__ __volatile__ ("rdtsc\n" : "=A" (ts));
- return ts;
-}
-
-#define OIL_STAMP oil_profile_stamp_rdtsc
-
-typedef struct pixman_timer_t pixman_timer_t;
-
-struct pixman_timer_t
-{
- int initialized;
- const char * name;
- uint64_t n_times;
- uint64_t total;
- pixman_timer_t *next;
-};
-
-extern int timer_defined;
-
-void pixman_timer_register (pixman_timer_t *timer);
-
-#define TIMER_BEGIN(tname) \
- { \
- static pixman_timer_t timer ## tname; \
- uint64_t begin ## tname; \
- \
- if (!timer ## tname.initialized) \
- { \
- timer ## tname.initialized = 1; \
- timer ## tname.name = # tname; \
- pixman_timer_register (&timer ## tname); \
- } \
- \
- timer ## tname.n_times++; \
- begin ## tname = OIL_STAMP ();
-
-#define TIMER_END(tname) \
- timer ## tname.total += OIL_STAMP () - begin ## tname; \
- }
-
-#endif /* PIXMAN_TIMERS */
-
-#endif /* PIXMAN_PRIVATE_H */
+#ifndef PACKAGE +# error config.h must be included before pixman-private.h +#endif + +#ifndef PIXMAN_PRIVATE_H +#define PIXMAN_PRIVATE_H + +#define PIXMAN_DISABLE_DEPRECATED +#define PIXMAN_USE_INTERNAL_API + +#include "pixman.h" +#include <time.h> +#include <assert.h> +#include <stdio.h> +#include <string.h> + +#include "pixman-compiler.h" + +/* + * Images + */ +typedef struct image_common image_common_t; +typedef struct solid_fill solid_fill_t; +typedef struct gradient gradient_t; +typedef struct linear_gradient linear_gradient_t; +typedef struct horizontal_gradient horizontal_gradient_t; +typedef struct vertical_gradient vertical_gradient_t; +typedef struct conical_gradient conical_gradient_t; +typedef struct radial_gradient radial_gradient_t; +typedef struct bits_image bits_image_t; +typedef struct circle circle_t; + +typedef void (*fetch_scanline_t) (pixman_image_t *image, + int x, + int y, + int width, + uint32_t *buffer, + const uint32_t *mask); + +typedef uint32_t (*fetch_pixel_32_t) (bits_image_t *image, + int x, + int y); + +typedef uint64_t (*fetch_pixel_64_t) (bits_image_t *image, + int x, + int y); + +typedef void (*store_scanline_t) (bits_image_t * image, + int x, + int y, + int width, + const uint32_t *values); + +typedef enum +{ + BITS, + LINEAR, + CONICAL, + RADIAL, + SOLID +} image_type_t; + +typedef void (*property_changed_func_t) (pixman_image_t *image); + +struct image_common +{ + image_type_t type; + int32_t ref_count; + pixman_region32_t clip_region; + int32_t alpha_count; /* How many times this image is being used as an alpha map */ + pixman_bool_t have_clip_region; /* FALSE if there is no clip */ + pixman_bool_t client_clip; /* Whether the source clip was + set by a client */ + pixman_bool_t clip_sources; /* Whether the clip applies when + * the image is used as a source + */ + pixman_bool_t dirty; + pixman_transform_t * transform; + pixman_repeat_t repeat; + pixman_filter_t filter; + pixman_fixed_t * filter_params; + int n_filter_params; + bits_image_t * alpha_map; + int alpha_origin_x; + int alpha_origin_y; + pixman_bool_t component_alpha; + property_changed_func_t property_changed; + + pixman_image_destroy_func_t destroy_func; + void * destroy_data; + + uint32_t flags; + pixman_format_code_t extended_format_code; +}; + +struct solid_fill +{ + image_common_t common; + pixman_color_t color; + + uint32_t color_32; + uint64_t color_64; +}; + +struct gradient +{ + image_common_t common; + int n_stops; + pixman_gradient_stop_t *stops; +}; + +struct linear_gradient +{ + gradient_t common; + pixman_point_fixed_t p1; + pixman_point_fixed_t p2; +}; + +struct circle +{ + pixman_fixed_t x; + pixman_fixed_t y; + pixman_fixed_t radius; +}; + +struct radial_gradient +{ + gradient_t common; + + circle_t c1; + circle_t c2; + + circle_t delta; + double a; + double inva; + double mindr; +}; + +struct conical_gradient +{ + gradient_t common; + pixman_point_fixed_t center; + double angle; +}; + +struct bits_image +{ + image_common_t common; + pixman_format_code_t format; + const pixman_indexed_t * indexed; + int width; + int height; + uint32_t * bits; + uint32_t * free_me; + int rowstride; /* in number of uint32_t's */ + + fetch_scanline_t get_scanline_32; + fetch_scanline_t get_scanline_64; + + fetch_scanline_t fetch_scanline_32; + fetch_pixel_32_t fetch_pixel_32; + store_scanline_t store_scanline_32; + + fetch_scanline_t fetch_scanline_64; + fetch_pixel_64_t fetch_pixel_64; + store_scanline_t store_scanline_64; + + /* Used for indirect access to the bits */ + pixman_read_memory_func_t read_func; + pixman_write_memory_func_t write_func; +}; + +union pixman_image +{ + image_type_t type; + image_common_t common; + bits_image_t bits; + gradient_t gradient; + linear_gradient_t linear; + conical_gradient_t conical; + radial_gradient_t radial; + solid_fill_t solid; +}; + +typedef struct pixman_iter_t pixman_iter_t; +typedef enum +{ + ITER_NARROW = (1 << 0), + + /* "Localized alpha" is when the alpha channel is used only to compute + * the alpha value of the destination. This means that the computation + * of the RGB values of the result is independent of the alpha value. + * + * For example, the OVER operator has localized alpha for the + * destination, because the RGB values of the result can be computed + * without knowing the destination alpha. Similarly, ADD has localized + * alpha for both source and destination because the RGB values of the + * result can be computed without knowing the alpha value of source or + * destination. + * + * When he destination is xRGB, this is useful knowledge, because then + * we can treat it as if it were ARGB, which means in some cases we can + * avoid copying it to a temporary buffer. + */ + ITER_LOCALIZED_ALPHA = (1 << 1), + ITER_IGNORE_ALPHA = (1 << 2), + ITER_IGNORE_RGB = (1 << 3) +} iter_flags_t; + +struct pixman_iter_t +{ + uint32_t *(* get_scanline) (pixman_iter_t *iter, const uint32_t *mask); + void (* write_back) (pixman_iter_t *iter); + + pixman_image_t * image; + uint32_t * buffer; + int x, y; + int width; +}; + +void +_pixman_bits_image_setup_accessors (bits_image_t *image); + +void +_pixman_bits_image_src_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags); +void +_pixman_bits_image_dest_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags); + +void +_pixman_solid_fill_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags); + +void +_pixman_linear_gradient_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags); + +void +_pixman_radial_gradient_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags); + +void +_pixman_conical_gradient_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags); + +pixman_image_t * +_pixman_image_allocate (void); + +pixman_bool_t +_pixman_init_gradient (gradient_t * gradient, + const pixman_gradient_stop_t *stops, + int n_stops); +void +_pixman_image_reset_clip_region (pixman_image_t *image); + +void +_pixman_image_validate (pixman_image_t *image); + +#define PIXMAN_IMAGE_GET_LINE(image, x, y, type, out_stride, line, mul) \ + do \ + { \ + uint32_t *__bits__; \ + int __stride__; \ + \ + __bits__ = image->bits.bits; \ + __stride__ = image->bits.rowstride; \ + (out_stride) = \ + __stride__ * (int) sizeof (uint32_t) / (int) sizeof (type); \ + (line) = \ + ((type *) __bits__) + (out_stride) * (y) + (mul) * (x); \ + } while (0) + +/* + * Gradient walker + */ +typedef struct +{ + uint32_t left_ag; + uint32_t left_rb; + uint32_t right_ag; + uint32_t right_rb; + int32_t left_x; + int32_t right_x; + int32_t stepper; + + pixman_gradient_stop_t *stops; + int num_stops; + unsigned int spread; + + int need_reset; +} pixman_gradient_walker_t; + +void +_pixman_gradient_walker_init (pixman_gradient_walker_t *walker, + gradient_t * gradient, + unsigned int spread); + +void +_pixman_gradient_walker_reset (pixman_gradient_walker_t *walker, + pixman_fixed_32_32_t pos); + +uint32_t +_pixman_gradient_walker_pixel (pixman_gradient_walker_t *walker, + pixman_fixed_32_32_t x); + +/* + * Edges + */ + +#define MAX_ALPHA(n) ((1 << (n)) - 1) +#define N_Y_FRAC(n) ((n) == 1 ? 1 : (1 << ((n) / 2)) - 1) +#define N_X_FRAC(n) ((n) == 1 ? 1 : (1 << ((n) / 2)) + 1) + +#define STEP_Y_SMALL(n) (pixman_fixed_1 / N_Y_FRAC (n)) +#define STEP_Y_BIG(n) (pixman_fixed_1 - (N_Y_FRAC (n) - 1) * STEP_Y_SMALL (n)) + +#define Y_FRAC_FIRST(n) (STEP_Y_BIG (n) / 2) +#define Y_FRAC_LAST(n) (Y_FRAC_FIRST (n) + (N_Y_FRAC (n) - 1) * STEP_Y_SMALL (n)) + +#define STEP_X_SMALL(n) (pixman_fixed_1 / N_X_FRAC (n)) +#define STEP_X_BIG(n) (pixman_fixed_1 - (N_X_FRAC (n) - 1) * STEP_X_SMALL (n)) + +#define X_FRAC_FIRST(n) (STEP_X_BIG (n) / 2) +#define X_FRAC_LAST(n) (X_FRAC_FIRST (n) + (N_X_FRAC (n) - 1) * STEP_X_SMALL (n)) + +#define RENDER_SAMPLES_X(x, n) \ + ((n) == 1? 0 : (pixman_fixed_frac (x) + \ + X_FRAC_FIRST (n)) / STEP_X_SMALL (n)) + +void +pixman_rasterize_edges_accessors (pixman_image_t *image, + pixman_edge_t * l, + pixman_edge_t * r, + pixman_fixed_t t, + pixman_fixed_t b); + +/* + * Implementations + */ +typedef struct pixman_implementation_t pixman_implementation_t; + +typedef void (*pixman_combine_32_func_t) (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width); + +typedef void (*pixman_combine_64_func_t) (pixman_implementation_t *imp, + pixman_op_t op, + uint64_t * dest, + const uint64_t * src, + const uint64_t * mask, + int width); + +typedef void (*pixman_composite_func_t) (pixman_implementation_t *imp, + pixman_op_t op, + pixman_image_t * src, + pixman_image_t * mask, + pixman_image_t * dest, + int32_t src_x, + int32_t src_y, + int32_t mask_x, + int32_t mask_y, + int32_t dest_x, + int32_t dest_y, + int32_t width, + int32_t height); +typedef pixman_bool_t (*pixman_blt_func_t) (pixman_implementation_t *imp, + uint32_t * src_bits, + uint32_t * dst_bits, + int src_stride, + int dst_stride, + int src_bpp, + int dst_bpp, + int src_x, + int src_y, + int dst_x, + int dst_y, + int width, + int height); +typedef pixman_bool_t (*pixman_fill_func_t) (pixman_implementation_t *imp, + uint32_t * bits, + int stride, + int bpp, + int x, + int y, + int width, + int height, + uint32_t xor); +typedef void (*pixman_iter_init_func_t) (pixman_implementation_t *imp, + pixman_iter_t *iter, + pixman_image_t *image, + int x, + int y, + int width, + int height, + uint8_t *buffer, + iter_flags_t flags); + +void _pixman_setup_combiner_functions_32 (pixman_implementation_t *imp); +void _pixman_setup_combiner_functions_64 (pixman_implementation_t *imp); + +typedef struct +{ + pixman_op_t op; + pixman_format_code_t src_format; + uint32_t src_flags; + pixman_format_code_t mask_format; + uint32_t mask_flags; + pixman_format_code_t dest_format; + uint32_t dest_flags; + pixman_composite_func_t func; +} pixman_fast_path_t; + +struct pixman_implementation_t +{ + pixman_implementation_t * toplevel; + pixman_implementation_t * delegate; + const pixman_fast_path_t * fast_paths; + + pixman_blt_func_t blt; + pixman_fill_func_t fill; + pixman_iter_init_func_t src_iter_init; + pixman_iter_init_func_t dest_iter_init; + + pixman_combine_32_func_t combine_32[PIXMAN_N_OPERATORS]; + pixman_combine_32_func_t combine_32_ca[PIXMAN_N_OPERATORS]; + pixman_combine_64_func_t combine_64[PIXMAN_N_OPERATORS]; + pixman_combine_64_func_t combine_64_ca[PIXMAN_N_OPERATORS]; +}; + +uint32_t +_pixman_image_get_solid (pixman_implementation_t *imp, + pixman_image_t * image, + pixman_format_code_t format); + +pixman_implementation_t * +_pixman_implementation_create (pixman_implementation_t *delegate, + const pixman_fast_path_t *fast_paths); + +void +_pixman_implementation_combine_32 (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width); +void +_pixman_implementation_combine_64 (pixman_implementation_t *imp, + pixman_op_t op, + uint64_t * dest, + const uint64_t * src, + const uint64_t * mask, + int width); +void +_pixman_implementation_combine_32_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint32_t * dest, + const uint32_t * src, + const uint32_t * mask, + int width); +void +_pixman_implementation_combine_64_ca (pixman_implementation_t *imp, + pixman_op_t op, + uint64_t * dest, + const uint64_t * src, + const uint64_t * mask, + int width); + +pixman_bool_t +_pixman_implementation_blt (pixman_implementation_t *imp, + uint32_t * src_bits, + uint32_t * dst_bits, + int src_stride, + int dst_stride, + int src_bpp, + int dst_bpp, + int src_x, + int src_y, + int dst_x, + int dst_y, + int width, + int height); + +pixman_bool_t +_pixman_implementation_fill (pixman_implementation_t *imp, + uint32_t * bits, + int stride, + int bpp, + int x, + int y, + int width, + int height, + uint32_t xor); + +void +_pixman_implementation_src_iter_init (pixman_implementation_t *imp, + pixman_iter_t *iter, + pixman_image_t *image, + int x, + int y, + int width, + int height, + uint8_t *buffer, + iter_flags_t flags); + +void +_pixman_implementation_dest_iter_init (pixman_implementation_t *imp, + pixman_iter_t *iter, + pixman_image_t *image, + int x, + int y, + int width, + int height, + uint8_t *buffer, + iter_flags_t flags); + +/* Specific implementations */ +pixman_implementation_t * +_pixman_implementation_create_general (void); + +pixman_implementation_t * +_pixman_implementation_create_fast_path (void); + +#ifdef USE_MMX +pixman_implementation_t * +_pixman_implementation_create_mmx (void); +#endif + +#ifdef USE_SSE2 +pixman_implementation_t * +_pixman_implementation_create_sse2 (void); +#endif + +#ifdef USE_ARM_SIMD +pixman_implementation_t * +_pixman_implementation_create_arm_simd (void); +#endif + +#ifdef USE_ARM_NEON +pixman_implementation_t * +_pixman_implementation_create_arm_neon (void); +#endif + +#ifdef USE_VMX +pixman_implementation_t * +_pixman_implementation_create_vmx (void); +#endif + +pixman_implementation_t * +_pixman_choose_implementation (void); + + + +/* + * Utilities + */ +uint32_t * +_pixman_iter_get_scanline_noop (pixman_iter_t *iter, const uint32_t *mask); + +/* These "formats" all have depth 0, so they + * will never clash with any real ones + */ +#define PIXMAN_null PIXMAN_FORMAT (0, 0, 0, 0, 0, 0) +#define PIXMAN_solid PIXMAN_FORMAT (0, 1, 0, 0, 0, 0) +#define PIXMAN_pixbuf PIXMAN_FORMAT (0, 2, 0, 0, 0, 0) +#define PIXMAN_rpixbuf PIXMAN_FORMAT (0, 3, 0, 0, 0, 0) +#define PIXMAN_unknown PIXMAN_FORMAT (0, 4, 0, 0, 0, 0) +#define PIXMAN_any PIXMAN_FORMAT (0, 5, 0, 0, 0, 0) + +#define PIXMAN_OP_any (PIXMAN_N_OPERATORS + 1) + +#define FAST_PATH_ID_TRANSFORM (1 << 0) +#define FAST_PATH_NO_ALPHA_MAP (1 << 1) +#define FAST_PATH_NO_CONVOLUTION_FILTER (1 << 2) +#define FAST_PATH_NO_PAD_REPEAT (1 << 3) +#define FAST_PATH_NO_REFLECT_REPEAT (1 << 4) +#define FAST_PATH_NO_ACCESSORS (1 << 5) +#define FAST_PATH_NARROW_FORMAT (1 << 6) +#define FAST_PATH_COMPONENT_ALPHA (1 << 8) +#define FAST_PATH_SAMPLES_OPAQUE (1 << 7) +#define FAST_PATH_UNIFIED_ALPHA (1 << 9) +#define FAST_PATH_SCALE_TRANSFORM (1 << 10) +#define FAST_PATH_NEAREST_FILTER (1 << 11) +#define FAST_PATH_HAS_TRANSFORM (1 << 12) +#define FAST_PATH_IS_OPAQUE (1 << 13) +#define FAST_PATH_NO_NORMAL_REPEAT (1 << 14) +#define FAST_PATH_NO_NONE_REPEAT (1 << 15) +#define FAST_PATH_SAMPLES_COVER_CLIP (1 << 16) +#define FAST_PATH_X_UNIT_POSITIVE (1 << 17) +#define FAST_PATH_AFFINE_TRANSFORM (1 << 18) +#define FAST_PATH_Y_UNIT_ZERO (1 << 19) +#define FAST_PATH_BILINEAR_FILTER (1 << 20) + +#define FAST_PATH_PAD_REPEAT \ + (FAST_PATH_NO_NONE_REPEAT | \ + FAST_PATH_NO_NORMAL_REPEAT | \ + FAST_PATH_NO_REFLECT_REPEAT) + +#define FAST_PATH_NORMAL_REPEAT \ + (FAST_PATH_NO_NONE_REPEAT | \ + FAST_PATH_NO_PAD_REPEAT | \ + FAST_PATH_NO_REFLECT_REPEAT) + +#define FAST_PATH_NONE_REPEAT \ + (FAST_PATH_NO_NORMAL_REPEAT | \ + FAST_PATH_NO_PAD_REPEAT | \ + FAST_PATH_NO_REFLECT_REPEAT) + +#define FAST_PATH_REFLECT_REPEAT \ + (FAST_PATH_NO_NONE_REPEAT | \ + FAST_PATH_NO_NORMAL_REPEAT | \ + FAST_PATH_NO_PAD_REPEAT) + +#define FAST_PATH_STANDARD_FLAGS \ + (FAST_PATH_NO_CONVOLUTION_FILTER | \ + FAST_PATH_NO_ACCESSORS | \ + FAST_PATH_NO_ALPHA_MAP | \ + FAST_PATH_NARROW_FORMAT) + +#define FAST_PATH_STD_DEST_FLAGS \ + (FAST_PATH_NO_ACCESSORS | \ + FAST_PATH_NO_ALPHA_MAP | \ + FAST_PATH_NARROW_FORMAT) + +#define SOURCE_FLAGS(format) \ + (FAST_PATH_STANDARD_FLAGS | \ + ((PIXMAN_ ## format == PIXMAN_solid) ? \ + 0 : (FAST_PATH_SAMPLES_COVER_CLIP | FAST_PATH_ID_TRANSFORM))) + +#define MASK_FLAGS(format, extra) \ + ((PIXMAN_ ## format == PIXMAN_null) ? 0 : (SOURCE_FLAGS (format) | extra)) + +#define FAST_PATH(op, src, src_flags, mask, mask_flags, dest, dest_flags, func) \ + PIXMAN_OP_ ## op, \ + PIXMAN_ ## src, \ + src_flags, \ + PIXMAN_ ## mask, \ + mask_flags, \ + PIXMAN_ ## dest, \ + dest_flags, \ + func + +#define PIXMAN_STD_FAST_PATH(op, src, mask, dest, func) \ + { FAST_PATH ( \ + op, \ + src, SOURCE_FLAGS (src), \ + mask, MASK_FLAGS (mask, FAST_PATH_UNIFIED_ALPHA), \ + dest, FAST_PATH_STD_DEST_FLAGS, \ + func) } + +#define PIXMAN_STD_FAST_PATH_CA(op, src, mask, dest, func) \ + { FAST_PATH ( \ + op, \ + src, SOURCE_FLAGS (src), \ + mask, MASK_FLAGS (mask, FAST_PATH_COMPONENT_ALPHA), \ + dest, FAST_PATH_STD_DEST_FLAGS, \ + func) } + +/* Memory allocation helpers */ +void * +pixman_malloc_ab (unsigned int n, unsigned int b); + +void * +pixman_malloc_abc (unsigned int a, unsigned int b, unsigned int c); + +pixman_bool_t +pixman_multiply_overflows_int (unsigned int a, unsigned int b); + +pixman_bool_t +pixman_addition_overflows_int (unsigned int a, unsigned int b); + +/* Compositing utilities */ +void +pixman_expand (uint64_t * dst, + const uint32_t * src, + pixman_format_code_t format, + int width); + +void +pixman_contract (uint32_t * dst, + const uint64_t *src, + int width); + + +/* Region Helpers */ +pixman_bool_t +pixman_region32_copy_from_region16 (pixman_region32_t *dst, + pixman_region16_t *src); + +pixman_bool_t +pixman_region16_copy_from_region32 (pixman_region16_t *dst, + pixman_region32_t *src); + + +/* Misc macros */ + +#ifndef FALSE +# define FALSE 0 +#endif + +#ifndef TRUE +# define TRUE 1 +#endif + +#ifndef MIN +# define MIN(a, b) ((a < b) ? a : b) +#endif + +#ifndef MAX +# define MAX(a, b) ((a > b) ? a : b) +#endif + +/* Integer division that rounds towards -infinity */ +#define DIV(a, b) \ + ((((a) < 0) == ((b) < 0)) ? (a) / (b) : \ + ((a) - (b) + 1 - (((b) < 0) << 1)) / (b)) + +/* Modulus that produces the remainder wrt. DIV */ +#define MOD(a, b) ((a) < 0 ? ((b) - ((-(a) - 1) % (b))) - 1 : (a) % (b)) + +#define CLIP(v, low, high) ((v) < (low) ? (low) : ((v) > (high) ? (high) : (v))) + +/* Conversion between 8888 and 0565 */ + +#define CONVERT_8888_TO_0565(s) \ + ((((s) >> 3) & 0x001f) | \ + (((s) >> 5) & 0x07e0) | \ + (((s) >> 8) & 0xf800)) + +#define CONVERT_0565_TO_0888(s) \ + (((((s) << 3) & 0xf8) | (((s) >> 2) & 0x7)) | \ + ((((s) << 5) & 0xfc00) | (((s) >> 1) & 0x300)) | \ + ((((s) << 8) & 0xf80000) | (((s) << 3) & 0x70000))) + +#define CONVERT_0565_TO_8888(s) (CONVERT_0565_TO_0888(s) | 0xff000000) + +/* Trivial versions that are useful in macros */ +#define CONVERT_8888_TO_8888(s) (s) +#define CONVERT_0565_TO_0565(s) (s) + +#define PIXMAN_FORMAT_IS_WIDE(f) \ + (PIXMAN_FORMAT_A (f) > 8 || \ + PIXMAN_FORMAT_R (f) > 8 || \ + PIXMAN_FORMAT_G (f) > 8 || \ + PIXMAN_FORMAT_B (f) > 8) + +#ifdef WORDS_BIGENDIAN +# define SCREEN_SHIFT_LEFT(x,n) ((x) << (n)) +# define SCREEN_SHIFT_RIGHT(x,n) ((x) >> (n)) +#else +# define SCREEN_SHIFT_LEFT(x,n) ((x) >> (n)) +# define SCREEN_SHIFT_RIGHT(x,n) ((x) << (n)) +#endif + +/* + * Various debugging code + */ + +#undef DEBUG + +#define COMPILE_TIME_ASSERT(x) \ + do { typedef int compile_time_assertion [(x)?1:-1]; } while (0) + +/* Turn on debugging depending on what type of release this is + */ +#if (((PIXMAN_VERSION_MICRO % 2) == 0) && ((PIXMAN_VERSION_MINOR % 2) == 1)) + +/* Debugging gets turned on for development releases because these + * are the things that end up in bleeding edge distributions such + * as Rawhide etc. + * + * For performance reasons we don't turn it on for stable releases or + * random git checkouts. (Random git checkouts are often used for + * performance work). + */ + +# define DEBUG + +#endif + +#ifdef DEBUG + +void +_pixman_log_error (const char *function, const char *message); + +#define return_if_fail(expr) \ + do \ + { \ + if (!(expr)) \ + { \ + _pixman_log_error (FUNC, "The expression " # expr " was false"); \ + return; \ + } \ + } \ + while (0) + +#define return_val_if_fail(expr, retval) \ + do \ + { \ + if (!(expr)) \ + { \ + _pixman_log_error (FUNC, "The expression " # expr " was false"); \ + return (retval); \ + } \ + } \ + while (0) + +#define critical_if_fail(expr) \ + do \ + { \ + if (!(expr)) \ + _pixman_log_error (FUNC, "The expression " # expr " was false"); \ + } \ + while (0) + + +#else + +#define _pixman_log_error(f,m) do { } while (0) \ + +#define return_if_fail(expr) \ + do \ + { \ + if (!(expr)) \ + return; \ + } \ + while (0) + +#define return_val_if_fail(expr, retval) \ + do \ + { \ + if (!(expr)) \ + return (retval); \ + } \ + while (0) + +#define critical_if_fail(expr) \ + do \ + { \ + } \ + while (0) +#endif + +/* + * Timers + */ + +#ifdef PIXMAN_TIMERS + +static inline uint64_t +oil_profile_stamp_rdtsc (void) +{ + uint64_t ts; + + __asm__ __volatile__ ("rdtsc\n" : "=A" (ts)); + return ts; +} + +#define OIL_STAMP oil_profile_stamp_rdtsc + +typedef struct pixman_timer_t pixman_timer_t; + +struct pixman_timer_t +{ + int initialized; + const char * name; + uint64_t n_times; + uint64_t total; + pixman_timer_t *next; +}; + +extern int timer_defined; + +void pixman_timer_register (pixman_timer_t *timer); + +#define TIMER_BEGIN(tname) \ + { \ + static pixman_timer_t timer ## tname; \ + uint64_t begin ## tname; \ + \ + if (!timer ## tname.initialized) \ + { \ + timer ## tname.initialized = 1; \ + timer ## tname.name = # tname; \ + pixman_timer_register (&timer ## tname); \ + } \ + \ + timer ## tname.n_times++; \ + begin ## tname = OIL_STAMP (); + +#define TIMER_END(tname) \ + timer ## tname.total += OIL_STAMP () - begin ## tname; \ + } + +#endif /* PIXMAN_TIMERS */ + +#endif /* PIXMAN_PRIVATE_H */ diff --git a/pixman/pixman/pixman-radial-gradient.c b/pixman/pixman/pixman-radial-gradient.c index 97c151865..6523b8259 100644 --- a/pixman/pixman/pixman-radial-gradient.c +++ b/pixman/pixman/pixman-radial-gradient.c @@ -1,447 +1,463 @@ -/* -*- Mode: c; c-basic-offset: 4; tab-width: 8; indent-tabs-mode: t; -*- */
-/*
- *
- * Copyright © 2000 Keith Packard, member of The XFree86 Project, Inc.
- * Copyright © 2000 SuSE, Inc.
- * 2005 Lars Knoll & Zack Rusin, Trolltech
- * Copyright © 2007 Red Hat, Inc.
- *
- *
- * Permission to use, copy, modify, distribute, and sell this software and its
- * documentation for any purpose is hereby granted without fee, provided that
- * the above copyright notice appear in all copies and that both that
- * copyright notice and this permission notice appear in supporting
- * documentation, and that the name of Keith Packard not be used in
- * advertising or publicity pertaining to distribution of the software without
- * specific, written prior permission. Keith Packard makes no
- * representations about the suitability of this software for any purpose. It
- * is provided "as is" without express or implied warranty.
- *
- * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
- * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
- * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
- * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING
- * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
- * SOFTWARE.
- */
-
-#ifdef HAVE_CONFIG_H
-#include <config.h>
-#endif
-#include <stdlib.h>
-#include <math.h>
-#include "pixman-private.h"
-
-static inline pixman_fixed_32_32_t
-dot (pixman_fixed_48_16_t x1,
- pixman_fixed_48_16_t y1,
- pixman_fixed_48_16_t z1,
- pixman_fixed_48_16_t x2,
- pixman_fixed_48_16_t y2,
- pixman_fixed_48_16_t z2)
-{
- /*
- * Exact computation, assuming that the input values can
- * be represented as pixman_fixed_16_16_t
- */
- return x1 * x2 + y1 * y2 + z1 * z2;
-}
-
-static inline double
-fdot (double x1,
- double y1,
- double z1,
- double x2,
- double y2,
- double z2)
-{
- /*
- * Error can be unbound in some special cases.
- * Using clever dot product algorithms (for example compensated
- * dot product) would improve this but make the code much less
- * obvious
- */
- return x1 * x2 + y1 * y2 + z1 * z2;
-}
-
-static uint32_t
-radial_compute_color (double a,
- double b,
- double c,
- double inva,
- double dr,
- double mindr,
- pixman_gradient_walker_t *walker,
- pixman_repeat_t repeat)
-{
- /*
- * In this function error propagation can lead to bad results:
- * - det can have an unbound error (if b*b-a*c is very small),
- * potentially making it the opposite sign of what it should have been
- * (thus clearing a pixel that would have been colored or vice-versa)
- * or propagating the error to sqrtdet;
- * if det has the wrong sign or b is very small, this can lead to bad
- * results
- *
- * - the algorithm used to compute the solutions of the quadratic
- * equation is not numerically stable (but saves one division compared
- * to the numerically stable one);
- * this can be a problem if a*c is much smaller than b*b
- *
- * - the above problems are worse if a is small (as inva becomes bigger)
- */
- double det;
-
- if (a == 0)
- {
- double t;
-
- if (b == 0)
- return 0;
-
- t = pixman_fixed_1 / 2 * c / b;
- if (repeat == PIXMAN_REPEAT_NONE)
- {
- if (0 <= t && t <= pixman_fixed_1)
- return _pixman_gradient_walker_pixel (walker, t);
- }
- else
- {
- if (t * dr > mindr)
- return _pixman_gradient_walker_pixel (walker, t);
- }
-
- return 0;
- }
-
- det = fdot (b, a, 0, b, -c, 0);
- if (det >= 0)
- {
- double sqrtdet, t0, t1;
-
- sqrtdet = sqrt (det);
- t0 = (b + sqrtdet) * inva;
- t1 = (b - sqrtdet) * inva;
-
- if (repeat == PIXMAN_REPEAT_NONE)
- {
- if (0 <= t0 && t0 <= pixman_fixed_1)
- return _pixman_gradient_walker_pixel (walker, t0);
- else if (0 <= t1 && t1 <= pixman_fixed_1)
- return _pixman_gradient_walker_pixel (walker, t1);
- }
- else
- {
- if (t0 * dr > mindr)
- return _pixman_gradient_walker_pixel (walker, t0);
- else if (t1 * dr > mindr)
- return _pixman_gradient_walker_pixel (walker, t1);
- }
- }
-
- return 0;
-}
-
-static void
-radial_gradient_get_scanline_32 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *mask)
-{
- /*
- * Implementation of radial gradients following the PDF specification.
- * See section 8.7.4.5.4 Type 3 (Radial) Shadings of the PDF Reference
- * Manual (PDF 32000-1:2008 at the time of this writing).
- *
- * In the radial gradient problem we are given two circles (c₁,r₁) and
- * (c₂,r₂) that define the gradient itself.
- *
- * Mathematically the gradient can be defined as the family of circles
- *
- * ((1-t)·c₁ + t·(c₂), (1-t)·r₁ + t·r₂)
- *
- * excluding those circles whose radius would be < 0. When a point
- * belongs to more than one circle, the one with a bigger t is the only
- * one that contributes to its color. When a point does not belong
- * to any of the circles, it is transparent black, i.e. RGBA (0, 0, 0, 0).
- * Further limitations on the range of values for t are imposed when
- * the gradient is not repeated, namely t must belong to [0,1].
- *
- * The graphical result is the same as drawing the valid (radius > 0)
- * circles with increasing t in [-inf, +inf] (or in [0,1] if the gradient
- * is not repeated) using SOURCE operatior composition.
- *
- * It looks like a cone pointing towards the viewer if the ending circle
- * is smaller than the starting one, a cone pointing inside the page if
- * the starting circle is the smaller one and like a cylinder if they
- * have the same radius.
- *
- * What we actually do is, given the point whose color we are interested
- * in, compute the t values for that point, solving for t in:
- *
- * length((1-t)·c₁ + t·(c₂) - p) = (1-t)·r₁ + t·r₂
- *
- * Let's rewrite it in a simpler way, by defining some auxiliary
- * variables:
- *
- * cd = c₂ - c₁
- * pd = p - c₁
- * dr = r₂ - r₁
- * lenght(t·cd - pd) = r₁ + t·dr
- *
- * which actually means
- *
- * hypot(t·cdx - pdx, t·cdy - pdy) = r₁ + t·dr
- *
- * or
- *
- * ⎷((t·cdx - pdx)² + (t·cdy - pdy)²) = r₁ + t·dr.
- *
- * If we impose (as stated earlier) that r₁ + t·dr >= 0, it becomes:
- *
- * (t·cdx - pdx)² + (t·cdy - pdy)² = (r₁ + t·dr)²
- *
- * where we can actually expand the squares and solve for t:
- *
- * t²cdx² - 2t·cdx·pdx + pdx² + t²cdy² - 2t·cdy·pdy + pdy² =
- * = r₁² + 2·r₁·t·dr + t²·dr²
- *
- * (cdx² + cdy² - dr²)t² - 2(cdx·pdx + cdy·pdy + r₁·dr)t +
- * (pdx² + pdy² - r₁²) = 0
- *
- * A = cdx² + cdy² - dr²
- * B = pdx·cdx + pdy·cdy + r₁·dr
- * C = pdx² + pdy² - r₁²
- * At² - 2Bt + C = 0
- *
- * The solutions (unless the equation degenerates because of A = 0) are:
- *
- * t = (B ± ⎷(B² - A·C)) / A
- *
- * The solution we are going to prefer is the bigger one, unless the
- * radius associated to it is negative (or it falls outside the valid t
- * range).
- *
- * Additional observations (useful for optimizations):
- * A does not depend on p
- *
- * A < 0 <=> one of the two circles completely contains the other one
- * <=> for every p, the radiuses associated with the two t solutions
- * have opposite sign
- */
-
- gradient_t *gradient = (gradient_t *)image;
- radial_gradient_t *radial = (radial_gradient_t *)image;
- uint32_t *end = buffer + width;
- pixman_gradient_walker_t walker;
- pixman_vector_t v, unit;
-
- /* reference point is the center of the pixel */
- v.vector[0] = pixman_int_to_fixed (x) + pixman_fixed_1 / 2;
- v.vector[1] = pixman_int_to_fixed (y) + pixman_fixed_1 / 2;
- v.vector[2] = pixman_fixed_1;
-
- _pixman_gradient_walker_init (&walker, gradient, image->common.repeat);
-
- if (image->common.transform)
- {
- if (!pixman_transform_point_3d (image->common.transform, &v))
- return;
-
- unit.vector[0] = image->common.transform->matrix[0][0];
- unit.vector[1] = image->common.transform->matrix[1][0];
- unit.vector[2] = image->common.transform->matrix[2][0];
- }
- else
- {
- unit.vector[0] = pixman_fixed_1;
- unit.vector[1] = 0;
- unit.vector[2] = 0;
- }
-
- if (unit.vector[2] == 0 && v.vector[2] == pixman_fixed_1)
- {
- /*
- * Given:
- *
- * t = (B ± ⎷(B² - A·C)) / A
- *
- * where
- *
- * A = cdx² + cdy² - dr²
- * B = pdx·cdx + pdy·cdy + r₁·dr
- * C = pdx² + pdy² - r₁²
- * det = B² - A·C
- *
- * Since we have an affine transformation, we know that (pdx, pdy)
- * increase linearly with each pixel,
- *
- * pdx = pdx₀ + n·ux,
- * pdy = pdy₀ + n·uy,
- *
- * we can then express B, C and det through multiple differentiation.
- */
- pixman_fixed_32_32_t b, db, c, dc, ddc;
-
- /* warning: this computation may overflow */
- v.vector[0] -= radial->c1.x;
- v.vector[1] -= radial->c1.y;
-
- /*
- * B and C are computed and updated exactly.
- * If fdot was used instead of dot, in the worst case it would
- * lose 11 bits of precision in each of the multiplication and
- * summing up would zero out all the bit that were preserved,
- * thus making the result 0 instead of the correct one.
- * This would mean a worst case of unbound relative error or
- * about 2^10 absolute error
- */
- b = dot (v.vector[0], v.vector[1], radial->c1.radius,
- radial->delta.x, radial->delta.y, radial->delta.radius);
- db = dot (unit.vector[0], unit.vector[1], 0,
- radial->delta.x, radial->delta.y, 0);
-
- c = dot (v.vector[0], v.vector[1],
- -((pixman_fixed_48_16_t) radial->c1.radius),
- v.vector[0], v.vector[1], radial->c1.radius);
- dc = dot (2 * (pixman_fixed_48_16_t) v.vector[0] + unit.vector[0],
- 2 * (pixman_fixed_48_16_t) v.vector[1] + unit.vector[1],
- 0,
- unit.vector[0], unit.vector[1], 0);
- ddc = 2 * dot (unit.vector[0], unit.vector[1], 0,
- unit.vector[0], unit.vector[1], 0);
-
- while (buffer < end)
- {
- if (!mask || *mask++)
- {
- *buffer = radial_compute_color (radial->a, b, c,
- radial->inva,
- radial->delta.radius,
- radial->mindr,
- &walker,
- image->common.repeat);
- }
-
- b += db;
- c += dc;
- dc += ddc;
- ++buffer;
- }
- }
- else
- {
- /* projective */
- /* Warning:
- * error propagation guarantees are much looser than in the affine case
- */
- while (buffer < end)
- {
- if (!mask || *mask++)
- {
- if (v.vector[2] != 0)
- {
- double pdx, pdy, invv2, b, c;
-
- invv2 = 1. * pixman_fixed_1 / v.vector[2];
-
- pdx = v.vector[0] * invv2 - radial->c1.x;
- /* / pixman_fixed_1 */
-
- pdy = v.vector[1] * invv2 - radial->c1.y;
- /* / pixman_fixed_1 */
-
- b = fdot (pdx, pdy, radial->c1.radius,
- radial->delta.x, radial->delta.y,
- radial->delta.radius);
- /* / pixman_fixed_1 / pixman_fixed_1 */
-
- c = fdot (pdx, pdy, -radial->c1.radius,
- pdx, pdy, radial->c1.radius);
- /* / pixman_fixed_1 / pixman_fixed_1 */
-
- *buffer = radial_compute_color (radial->a, b, c,
- radial->inva,
- radial->delta.radius,
- radial->mindr,
- &walker,
- image->common.repeat);
- }
- else
- {
- *buffer = 0;
- }
- }
-
- ++buffer;
-
- v.vector[0] += unit.vector[0];
- v.vector[1] += unit.vector[1];
- v.vector[2] += unit.vector[2];
- }
- }
-}
-
-static void
-radial_gradient_property_changed (pixman_image_t *image)
-{
- image->common.get_scanline_32 = radial_gradient_get_scanline_32;
- image->common.get_scanline_64 = _pixman_image_get_scanline_generic_64;
-}
-
-PIXMAN_EXPORT pixman_image_t *
-pixman_image_create_radial_gradient (pixman_point_fixed_t * inner,
- pixman_point_fixed_t * outer,
- pixman_fixed_t inner_radius,
- pixman_fixed_t outer_radius,
- const pixman_gradient_stop_t *stops,
- int n_stops)
-{
- pixman_image_t *image;
- radial_gradient_t *radial;
-
- image = _pixman_image_allocate ();
-
- if (!image)
- return NULL;
-
- radial = &image->radial;
-
- if (!_pixman_init_gradient (&radial->common, stops, n_stops))
- {
- free (image);
- return NULL;
- }
-
- image->type = RADIAL;
-
- radial->c1.x = inner->x;
- radial->c1.y = inner->y;
- radial->c1.radius = inner_radius;
- radial->c2.x = outer->x;
- radial->c2.y = outer->y;
- radial->c2.radius = outer_radius;
-
- /* warning: this computations may overflow */
- radial->delta.x = radial->c2.x - radial->c1.x;
- radial->delta.y = radial->c2.y - radial->c1.y;
- radial->delta.radius = radial->c2.radius - radial->c1.radius;
-
- /* computed exactly, then cast to double -> every bit of the double
- representation is correct (53 bits) */
- radial->a = dot (radial->delta.x, radial->delta.y, -radial->delta.radius,
- radial->delta.x, radial->delta.y, radial->delta.radius);
- if (radial->a != 0)
- radial->inva = 1. * pixman_fixed_1 / radial->a;
-
- radial->mindr = -1. * pixman_fixed_1 * radial->c1.radius;
-
- image->common.property_changed = radial_gradient_property_changed;
-
- return image;
-}
-
+/* -*- Mode: c; c-basic-offset: 4; tab-width: 8; indent-tabs-mode: t; -*- */ +/* + * + * Copyright © 2000 Keith Packard, member of The XFree86 Project, Inc. + * Copyright © 2000 SuSE, Inc. + * 2005 Lars Knoll & Zack Rusin, Trolltech + * Copyright © 2007 Red Hat, Inc. + * + * + * Permission to use, copy, modify, distribute, and sell this software and its + * documentation for any purpose is hereby granted without fee, provided that + * the above copyright notice appear in all copies and that both that + * copyright notice and this permission notice appear in supporting + * documentation, and that the name of Keith Packard not be used in + * advertising or publicity pertaining to distribution of the software without + * specific, written prior permission. Keith Packard makes no + * representations about the suitability of this software for any purpose. It + * is provided "as is" without express or implied warranty. + * + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS + * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND + * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY + * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN + * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING + * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS + * SOFTWARE. + */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif +#include <stdlib.h> +#include <math.h> +#include "pixman-private.h" + +static inline pixman_fixed_32_32_t +dot (pixman_fixed_48_16_t x1, + pixman_fixed_48_16_t y1, + pixman_fixed_48_16_t z1, + pixman_fixed_48_16_t x2, + pixman_fixed_48_16_t y2, + pixman_fixed_48_16_t z2) +{ + /* + * Exact computation, assuming that the input values can + * be represented as pixman_fixed_16_16_t + */ + return x1 * x2 + y1 * y2 + z1 * z2; +} + +static inline double +fdot (double x1, + double y1, + double z1, + double x2, + double y2, + double z2) +{ + /* + * Error can be unbound in some special cases. + * Using clever dot product algorithms (for example compensated + * dot product) would improve this but make the code much less + * obvious + */ + return x1 * x2 + y1 * y2 + z1 * z2; +} + +static uint32_t +radial_compute_color (double a, + double b, + double c, + double inva, + double dr, + double mindr, + pixman_gradient_walker_t *walker, + pixman_repeat_t repeat) +{ + /* + * In this function error propagation can lead to bad results: + * - det can have an unbound error (if b*b-a*c is very small), + * potentially making it the opposite sign of what it should have been + * (thus clearing a pixel that would have been colored or vice-versa) + * or propagating the error to sqrtdet; + * if det has the wrong sign or b is very small, this can lead to bad + * results + * + * - the algorithm used to compute the solutions of the quadratic + * equation is not numerically stable (but saves one division compared + * to the numerically stable one); + * this can be a problem if a*c is much smaller than b*b + * + * - the above problems are worse if a is small (as inva becomes bigger) + */ + double det; + + if (a == 0) + { + double t; + + if (b == 0) + return 0; + + t = pixman_fixed_1 / 2 * c / b; + if (repeat == PIXMAN_REPEAT_NONE) + { + if (0 <= t && t <= pixman_fixed_1) + return _pixman_gradient_walker_pixel (walker, t); + } + else + { + if (t * dr > mindr) + return _pixman_gradient_walker_pixel (walker, t); + } + + return 0; + } + + det = fdot (b, a, 0, b, -c, 0); + if (det >= 0) + { + double sqrtdet, t0, t1; + + sqrtdet = sqrt (det); + t0 = (b + sqrtdet) * inva; + t1 = (b - sqrtdet) * inva; + + if (repeat == PIXMAN_REPEAT_NONE) + { + if (0 <= t0 && t0 <= pixman_fixed_1) + return _pixman_gradient_walker_pixel (walker, t0); + else if (0 <= t1 && t1 <= pixman_fixed_1) + return _pixman_gradient_walker_pixel (walker, t1); + } + else + { + if (t0 * dr > mindr) + return _pixman_gradient_walker_pixel (walker, t0); + else if (t1 * dr > mindr) + return _pixman_gradient_walker_pixel (walker, t1); + } + } + + return 0; +} + +static uint32_t * +radial_get_scanline_narrow (pixman_iter_t *iter, const uint32_t *mask) +{ + /* + * Implementation of radial gradients following the PDF specification. + * See section 8.7.4.5.4 Type 3 (Radial) Shadings of the PDF Reference + * Manual (PDF 32000-1:2008 at the time of this writing). + * + * In the radial gradient problem we are given two circles (c₁,r₁) and + * (c₂,r₂) that define the gradient itself. + * + * Mathematically the gradient can be defined as the family of circles + * + * ((1-t)·c₁ + t·(c₂), (1-t)·r₁ + t·r₂) + * + * excluding those circles whose radius would be < 0. When a point + * belongs to more than one circle, the one with a bigger t is the only + * one that contributes to its color. When a point does not belong + * to any of the circles, it is transparent black, i.e. RGBA (0, 0, 0, 0). + * Further limitations on the range of values for t are imposed when + * the gradient is not repeated, namely t must belong to [0,1]. + * + * The graphical result is the same as drawing the valid (radius > 0) + * circles with increasing t in [-inf, +inf] (or in [0,1] if the gradient + * is not repeated) using SOURCE operatior composition. + * + * It looks like a cone pointing towards the viewer if the ending circle + * is smaller than the starting one, a cone pointing inside the page if + * the starting circle is the smaller one and like a cylinder if they + * have the same radius. + * + * What we actually do is, given the point whose color we are interested + * in, compute the t values for that point, solving for t in: + * + * length((1-t)·c₁ + t·(c₂) - p) = (1-t)·r₁ + t·r₂ + * + * Let's rewrite it in a simpler way, by defining some auxiliary + * variables: + * + * cd = c₂ - c₁ + * pd = p - c₁ + * dr = r₂ - r₁ + * lenght(t·cd - pd) = r₁ + t·dr + * + * which actually means + * + * hypot(t·cdx - pdx, t·cdy - pdy) = r₁ + t·dr + * + * or + * + * ⎷((t·cdx - pdx)² + (t·cdy - pdy)²) = r₁ + t·dr. + * + * If we impose (as stated earlier) that r₁ + t·dr >= 0, it becomes: + * + * (t·cdx - pdx)² + (t·cdy - pdy)² = (r₁ + t·dr)² + * + * where we can actually expand the squares and solve for t: + * + * t²cdx² - 2t·cdx·pdx + pdx² + t²cdy² - 2t·cdy·pdy + pdy² = + * = r₁² + 2·r₁·t·dr + t²·dr² + * + * (cdx² + cdy² - dr²)t² - 2(cdx·pdx + cdy·pdy + r₁·dr)t + + * (pdx² + pdy² - r₁²) = 0 + * + * A = cdx² + cdy² - dr² + * B = pdx·cdx + pdy·cdy + r₁·dr + * C = pdx² + pdy² - r₁² + * At² - 2Bt + C = 0 + * + * The solutions (unless the equation degenerates because of A = 0) are: + * + * t = (B ± ⎷(B² - A·C)) / A + * + * The solution we are going to prefer is the bigger one, unless the + * radius associated to it is negative (or it falls outside the valid t + * range). + * + * Additional observations (useful for optimizations): + * A does not depend on p + * + * A < 0 <=> one of the two circles completely contains the other one + * <=> for every p, the radiuses associated with the two t solutions + * have opposite sign + */ + pixman_image_t *image = iter->image; + int x = iter->x; + int y = iter->y; + int width = iter->width; + uint32_t *buffer = iter->buffer; + + gradient_t *gradient = (gradient_t *)image; + radial_gradient_t *radial = (radial_gradient_t *)image; + uint32_t *end = buffer + width; + pixman_gradient_walker_t walker; + pixman_vector_t v, unit; + + /* reference point is the center of the pixel */ + v.vector[0] = pixman_int_to_fixed (x) + pixman_fixed_1 / 2; + v.vector[1] = pixman_int_to_fixed (y) + pixman_fixed_1 / 2; + v.vector[2] = pixman_fixed_1; + + _pixman_gradient_walker_init (&walker, gradient, image->common.repeat); + + if (image->common.transform) + { + if (!pixman_transform_point_3d (image->common.transform, &v)) + return iter->buffer; + + unit.vector[0] = image->common.transform->matrix[0][0]; + unit.vector[1] = image->common.transform->matrix[1][0]; + unit.vector[2] = image->common.transform->matrix[2][0]; + } + else + { + unit.vector[0] = pixman_fixed_1; + unit.vector[1] = 0; + unit.vector[2] = 0; + } + + if (unit.vector[2] == 0 && v.vector[2] == pixman_fixed_1) + { + /* + * Given: + * + * t = (B ± ⎷(B² - A·C)) / A + * + * where + * + * A = cdx² + cdy² - dr² + * B = pdx·cdx + pdy·cdy + r₁·dr + * C = pdx² + pdy² - r₁² + * det = B² - A·C + * + * Since we have an affine transformation, we know that (pdx, pdy) + * increase linearly with each pixel, + * + * pdx = pdx₀ + n·ux, + * pdy = pdy₀ + n·uy, + * + * we can then express B, C and det through multiple differentiation. + */ + pixman_fixed_32_32_t b, db, c, dc, ddc; + + /* warning: this computation may overflow */ + v.vector[0] -= radial->c1.x; + v.vector[1] -= radial->c1.y; + + /* + * B and C are computed and updated exactly. + * If fdot was used instead of dot, in the worst case it would + * lose 11 bits of precision in each of the multiplication and + * summing up would zero out all the bit that were preserved, + * thus making the result 0 instead of the correct one. + * This would mean a worst case of unbound relative error or + * about 2^10 absolute error + */ + b = dot (v.vector[0], v.vector[1], radial->c1.radius, + radial->delta.x, radial->delta.y, radial->delta.radius); + db = dot (unit.vector[0], unit.vector[1], 0, + radial->delta.x, radial->delta.y, 0); + + c = dot (v.vector[0], v.vector[1], + -((pixman_fixed_48_16_t) radial->c1.radius), + v.vector[0], v.vector[1], radial->c1.radius); + dc = dot (2 * (pixman_fixed_48_16_t) v.vector[0] + unit.vector[0], + 2 * (pixman_fixed_48_16_t) v.vector[1] + unit.vector[1], + 0, + unit.vector[0], unit.vector[1], 0); + ddc = 2 * dot (unit.vector[0], unit.vector[1], 0, + unit.vector[0], unit.vector[1], 0); + + while (buffer < end) + { + if (!mask || *mask++) + { + *buffer = radial_compute_color (radial->a, b, c, + radial->inva, + radial->delta.radius, + radial->mindr, + &walker, + image->common.repeat); + } + + b += db; + c += dc; + dc += ddc; + ++buffer; + } + } + else + { + /* projective */ + /* Warning: + * error propagation guarantees are much looser than in the affine case + */ + while (buffer < end) + { + if (!mask || *mask++) + { + if (v.vector[2] != 0) + { + double pdx, pdy, invv2, b, c; + + invv2 = 1. * pixman_fixed_1 / v.vector[2]; + + pdx = v.vector[0] * invv2 - radial->c1.x; + /* / pixman_fixed_1 */ + + pdy = v.vector[1] * invv2 - radial->c1.y; + /* / pixman_fixed_1 */ + + b = fdot (pdx, pdy, radial->c1.radius, + radial->delta.x, radial->delta.y, + radial->delta.radius); + /* / pixman_fixed_1 / pixman_fixed_1 */ + + c = fdot (pdx, pdy, -radial->c1.radius, + pdx, pdy, radial->c1.radius); + /* / pixman_fixed_1 / pixman_fixed_1 */ + + *buffer = radial_compute_color (radial->a, b, c, + radial->inva, + radial->delta.radius, + radial->mindr, + &walker, + image->common.repeat); + } + else + { + *buffer = 0; + } + } + + ++buffer; + + v.vector[0] += unit.vector[0]; + v.vector[1] += unit.vector[1]; + v.vector[2] += unit.vector[2]; + } + } + + iter->y++; + return iter->buffer; +} + +static uint32_t * +radial_get_scanline_wide (pixman_iter_t *iter, const uint32_t *mask) +{ + uint32_t *buffer = radial_get_scanline_narrow (iter, NULL); + + pixman_expand ((uint64_t *)buffer, buffer, PIXMAN_a8r8g8b8, iter->width); + + return buffer; +} + +void +_pixman_radial_gradient_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags) +{ + if (flags & ITER_NARROW) + iter->get_scanline = radial_get_scanline_narrow; + else + iter->get_scanline = radial_get_scanline_wide; +} + +PIXMAN_EXPORT pixman_image_t * +pixman_image_create_radial_gradient (pixman_point_fixed_t * inner, + pixman_point_fixed_t * outer, + pixman_fixed_t inner_radius, + pixman_fixed_t outer_radius, + const pixman_gradient_stop_t *stops, + int n_stops) +{ + pixman_image_t *image; + radial_gradient_t *radial; + + image = _pixman_image_allocate (); + + if (!image) + return NULL; + + radial = &image->radial; + + if (!_pixman_init_gradient (&radial->common, stops, n_stops)) + { + free (image); + return NULL; + } + + image->type = RADIAL; + + radial->c1.x = inner->x; + radial->c1.y = inner->y; + radial->c1.radius = inner_radius; + radial->c2.x = outer->x; + radial->c2.y = outer->y; + radial->c2.radius = outer_radius; + + /* warning: this computations may overflow */ + radial->delta.x = radial->c2.x - radial->c1.x; + radial->delta.y = radial->c2.y - radial->c1.y; + radial->delta.radius = radial->c2.radius - radial->c1.radius; + + /* computed exactly, then cast to double -> every bit of the double + representation is correct (53 bits) */ + radial->a = dot (radial->delta.x, radial->delta.y, -radial->delta.radius, + radial->delta.x, radial->delta.y, radial->delta.radius); + if (radial->a != 0) + radial->inva = 1. * pixman_fixed_1 / radial->a; + + radial->mindr = -1. * pixman_fixed_1 * radial->c1.radius; + + return image; +} + diff --git a/pixman/pixman/pixman-solid-fill.c b/pixman/pixman/pixman-solid-fill.c index afff3c479..67681f2c0 100644 --- a/pixman/pixman/pixman-solid-fill.c +++ b/pixman/pixman/pixman-solid-fill.c @@ -1,117 +1,92 @@ -/*
- * Copyright © 2000 SuSE, Inc.
- * Copyright © 2007, 2009 Red Hat, Inc.
- * Copyright © 2009 Soren Sandmann
- *
- * Permission to use, copy, modify, distribute, and sell this software and its
- * documentation for any purpose is hereby granted without fee, provided that
- * the above copyright notice appear in all copies and that both that
- * copyright notice and this permission notice appear in supporting
- * documentation, and that the name of SuSE not be used in advertising or
- * publicity pertaining to distribution of the software without specific,
- * written prior permission. SuSE makes no representations about the
- * suitability of this software for any purpose. It is provided "as is"
- * without express or implied warranty.
- *
- * SuSE DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL SuSE
- * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
- * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
- * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
- */
-
-#ifdef HAVE_CONFIG_H
-#include <config.h>
-#endif
-#include "pixman-private.h"
-
-static void
-solid_fill_get_scanline_32 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *mask)
-{
- uint32_t *end = buffer + width;
- uint32_t color = image->solid.color_32;
-
- while (buffer < end)
- *(buffer++) = color;
-
- return;
-}
-
-static void
-solid_fill_get_scanline_64 (pixman_image_t *image,
- int x,
- int y,
- int width,
- uint32_t * buffer,
- const uint32_t *mask)
-{
- uint64_t *b = (uint64_t *)buffer;
- uint64_t *e = b + width;
- uint64_t color = image->solid.color_64;
-
- while (b < e)
- *(b++) = color;
-}
-
-static source_image_class_t
-solid_fill_classify (pixman_image_t *image,
- int x,
- int y,
- int width,
- int height)
-{
- return SOURCE_IMAGE_CLASS_HORIZONTAL;
-}
-
-static void
-solid_fill_property_changed (pixman_image_t *image)
-{
- image->common.get_scanline_32 = solid_fill_get_scanline_32;
- image->common.get_scanline_64 = solid_fill_get_scanline_64;
-}
-
-static uint32_t
-color_to_uint32 (const pixman_color_t *color)
-{
- return
- (color->alpha >> 8 << 24) |
- (color->red >> 8 << 16) |
- (color->green & 0xff00) |
- (color->blue >> 8);
-}
-
-static uint64_t
-color_to_uint64 (const pixman_color_t *color)
-{
- return
- ((uint64_t)color->alpha << 48) |
- ((uint64_t)color->red << 32) |
- ((uint64_t)color->green << 16) |
- ((uint64_t)color->blue);
-}
-
-PIXMAN_EXPORT pixman_image_t *
-pixman_image_create_solid_fill (pixman_color_t *color)
-{
- pixman_image_t *img = _pixman_image_allocate ();
-
- if (!img)
- return NULL;
-
- img->type = SOLID;
- img->solid.color = *color;
- img->solid.color_32 = color_to_uint32 (color);
- img->solid.color_64 = color_to_uint64 (color);
-
- img->common.classify = solid_fill_classify;
- img->common.property_changed = solid_fill_property_changed;
-
- return img;
-}
-
+/* + * Copyright © 2000 SuSE, Inc. + * Copyright © 2007, 2009 Red Hat, Inc. + * Copyright © 2009 Soren Sandmann + * + * Permission to use, copy, modify, distribute, and sell this software and its + * documentation for any purpose is hereby granted without fee, provided that + * the above copyright notice appear in all copies and that both that + * copyright notice and this permission notice appear in supporting + * documentation, and that the name of SuSE not be used in advertising or + * publicity pertaining to distribution of the software without specific, + * written prior permission. SuSE makes no representations about the + * suitability of this software for any purpose. It is provided "as is" + * without express or implied warranty. + * + * SuSE DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL SuSE + * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION + * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN + * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. + */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif +#include "pixman-private.h" + +void +_pixman_solid_fill_iter_init (pixman_image_t *image, + pixman_iter_t *iter, + int x, int y, int width, int height, + uint8_t *buffer, iter_flags_t flags) +{ + if (flags & ITER_NARROW) + { + uint32_t *b = (uint32_t *)buffer; + uint32_t *e = b + width; + uint32_t color = image->solid.color_32; + + while (b < e) + *(b++) = color; + } + else + { + uint64_t *b = (uint64_t *)buffer; + uint64_t *e = b + width; + uint64_t color = image->solid.color_64; + + while (b < e) + *(b++) = color; + } + + iter->get_scanline = _pixman_iter_get_scanline_noop; +} + +static uint32_t +color_to_uint32 (const pixman_color_t *color) +{ + return + (color->alpha >> 8 << 24) | + (color->red >> 8 << 16) | + (color->green & 0xff00) | + (color->blue >> 8); +} + +static uint64_t +color_to_uint64 (const pixman_color_t *color) +{ + return + ((uint64_t)color->alpha << 48) | + ((uint64_t)color->red << 32) | + ((uint64_t)color->green << 16) | + ((uint64_t)color->blue); +} + +PIXMAN_EXPORT pixman_image_t * +pixman_image_create_solid_fill (pixman_color_t *color) +{ + pixman_image_t *img = _pixman_image_allocate (); + + if (!img) + return NULL; + + img->type = SOLID; + img->solid.color = *color; + img->solid.color_32 = color_to_uint32 (color); + img->solid.color_64 = color_to_uint64 (color); + + return img; +} + diff --git a/pixman/pixman/pixman-sse2.c b/pixman/pixman/pixman-sse2.c index 94ba54cbf..3c0a42f41 100644 --- a/pixman/pixman/pixman-sse2.c +++ b/pixman/pixman/pixman-sse2.c @@ -2598,7 +2598,7 @@ sse2_composite_over_n_8888 (pixman_implementation_t *imp, __m128i xmm_src, xmm_alpha; __m128i xmm_dst, xmm_dst_lo, xmm_dst_hi; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); if (src == 0) return; @@ -2681,7 +2681,7 @@ sse2_composite_over_n_0565 (pixman_implementation_t *imp, __m128i xmm_src, xmm_alpha; __m128i xmm_dst, xmm_dst0, xmm_dst1, xmm_dst2, xmm_dst3; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); if (src == 0) return; @@ -2776,7 +2776,7 @@ sse2_composite_add_n_8888_8888_ca (pixman_implementation_t *imp, __m64 mmx_src, mmx_alpha, mmx_mask, mmx_dest; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -2904,7 +2904,7 @@ sse2_composite_over_n_8888_8888_ca (pixman_implementation_t *imp, __m64 mmx_src, mmx_alpha, mmx_mask, mmx_dest; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); if (src == 0) return; @@ -3036,7 +3036,7 @@ sse2_composite_over_8888_n_8888 (pixman_implementation_t *imp, PIXMAN_IMAGE_GET_LINE ( src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); - mask = _pixman_image_get_solid (mask_image, PIXMAN_a8r8g8b8); + mask = _pixman_image_get_solid (imp, mask_image, PIXMAN_a8r8g8b8); xmm_mask = create_mask_16_128 (mask >> 24); @@ -3226,7 +3226,7 @@ sse2_composite_over_x888_n_8888 (pixman_implementation_t *imp, PIXMAN_IMAGE_GET_LINE ( src_image, src_x, src_y, uint32_t, src_stride, src_line, 1); - mask = _pixman_image_get_solid (mask_image, PIXMAN_a8r8g8b8); + mask = _pixman_image_get_solid (imp, mask_image, PIXMAN_a8r8g8b8); xmm_mask = create_mask_16_128 (mask >> 24); xmm_alpha = mask_00ff; @@ -3498,7 +3498,7 @@ sse2_composite_over_n_8_8888 (pixman_implementation_t *imp, __m64 mmx_src, mmx_alpha, mmx_mask, mmx_dest; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -3782,7 +3782,7 @@ sse2_composite_src_n_8_8888 (pixman_implementation_t *imp, __m128i xmm_src, xmm_def; __m128i xmm_mask, xmm_mask_lo, xmm_mask_hi; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -3918,7 +3918,7 @@ sse2_composite_over_n_8_0565 (pixman_implementation_t *imp, __m128i xmm_mask, xmm_mask_lo, xmm_mask_hi; __m128i xmm_dst, xmm_dst0, xmm_dst1, xmm_dst2, xmm_dst3; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); srca = src >> 24; if (src == 0) @@ -4318,7 +4318,7 @@ sse2_composite_over_n_8888_0565_ca (pixman_implementation_t *imp, __m64 mmx_src, mmx_alpha, mmx_mask, mmx_dest; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); if (src == 0) return; @@ -4471,7 +4471,7 @@ sse2_composite_in_n_8_8 (pixman_implementation_t *imp, PIXMAN_IMAGE_GET_LINE ( mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); sa = src >> 24; @@ -4570,7 +4570,7 @@ sse2_composite_in_n_8 (pixman_implementation_t *imp, PIXMAN_IMAGE_GET_LINE ( dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1); - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); xmm_alpha = expand_alpha_1x128 (expand_pixel_32_1x128 (src)); @@ -4758,7 +4758,7 @@ sse2_composite_add_n_8_8 (pixman_implementation_t *imp, PIXMAN_IMAGE_GET_LINE ( mask_image, mask_x, mask_y, uint8_t, mask_stride, mask_line, 1); - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); sa = src >> 24; @@ -4855,7 +4855,7 @@ sse2_composite_add_n_8 (pixman_implementation_t *imp, PIXMAN_IMAGE_GET_LINE ( dst_image, dest_x, dest_y, uint8_t, dst_stride, dst_line, 1); - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); src >>= 24; @@ -5480,7 +5480,7 @@ sse2_composite_over_reverse_n_8888 (pixman_implementation_t *imp, int dst_stride; int32_t w; - src = _pixman_image_get_solid (src_image, dst_image->bits.format); + src = _pixman_image_get_solid (imp, src_image, dst_image->bits.format); if (src == 0) return; diff --git a/pixman/pixman/pixman-utils.c b/pixman/pixman/pixman-utils.c index 3ef88b753..cb4e62199 100644 --- a/pixman/pixman/pixman-utils.c +++ b/pixman/pixman/pixman-utils.c @@ -167,6 +167,12 @@ pixman_contract (uint32_t * dst, } } +uint32_t * +_pixman_iter_get_scanline_noop (pixman_iter_t *iter, const uint32_t *mask) +{ + return iter->buffer; +} + #define N_TMP_BOXES (16) pixman_bool_t |