#include <emmintrin.h> /* for SSE2 intrinsics */
#include "pixman-private.h"
#include "pixman-combine32.h"
-#include "pixman-fast-path.h"
+#include "pixman-inlines.h"
static __m128i mask_0080;
static __m128i mask_00ff;
byte_line += stride;
w = byte_width;
- while (w >= 1 && ((unsigned long)d & 1))
+ if (w >= 1 && ((unsigned long)d & 1))
{
*(uint8_t *)d = data;
w -= 1;
scaled_nearest_scanline_sse2_8888_n_8888_OVER,
uint32_t, uint32_t, uint32_t, NONE, TRUE, TRUE)
-static void
-bilinear_interpolate_line_sse2 (uint32_t * out,
- const uint32_t * top,
- const uint32_t * bottom,
- int wt,
- int wb,
- pixman_fixed_t x,
- pixman_fixed_t ux,
- int width)
-{
- const __m128i xmm_wt = _mm_set_epi16 (wt, wt, wt, wt, wt, wt, wt, wt);
- const __m128i xmm_wb = _mm_set_epi16 (wb, wb, wb, wb, wb, wb, wb, wb);
- const __m128i xmm_xorc = _mm_set_epi16 (0, 0, 0, 0, 0xff, 0xff, 0xff, 0xff);
- const __m128i xmm_addc = _mm_set_epi16 (0, 0, 0, 0, 1, 1, 1, 1);
- const __m128i xmm_ux = _mm_set_epi16 (ux, ux, ux, ux, ux, ux, ux, ux);
- const __m128i xmm_zero = _mm_setzero_si128 ();
- __m128i xmm_x = _mm_set_epi16 (x, x, x, x, x, x, x, x);
- uint32_t pix1, pix2, pix3, pix4;
-
- #define INTERPOLATE_ONE_PIXEL(pix) \
- do { \
- __m128i xmm_wh, xmm_lo, xmm_hi, a; \
- /* fetch 2x2 pixel block into sse2 register */ \
- uint32_t tl = top [pixman_fixed_to_int (x)]; \
- uint32_t tr = top [pixman_fixed_to_int (x) + 1]; \
- uint32_t bl = bottom [pixman_fixed_to_int (x)]; \
- uint32_t br = bottom [pixman_fixed_to_int (x) + 1]; \
- a = _mm_set_epi32 (tr, tl, br, bl); \
- x += ux; \
- /* vertical interpolation */ \
- a = _mm_add_epi16 (_mm_mullo_epi16 (_mm_unpackhi_epi8 (a, xmm_zero), \
- xmm_wt), \
- _mm_mullo_epi16 (_mm_unpacklo_epi8 (a, xmm_zero), \
- xmm_wb)); \
- /* calculate horizontal weights */ \
- xmm_wh = _mm_add_epi16 (xmm_addc, \
- _mm_xor_si128 (xmm_xorc, \
- _mm_srli_epi16 (xmm_x, 8))); \
- xmm_x = _mm_add_epi16 (xmm_x, xmm_ux); \
- /* horizontal interpolation */ \
- xmm_lo = _mm_mullo_epi16 (a, xmm_wh); \
- xmm_hi = _mm_mulhi_epu16 (a, xmm_wh); \
- a = _mm_add_epi32 (_mm_unpacklo_epi16 (xmm_lo, xmm_hi), \
- _mm_unpackhi_epi16 (xmm_lo, xmm_hi)); \
- /* shift and pack the result */ \
- a = _mm_srli_epi32 (a, 16); \
- a = _mm_packs_epi32 (a, a); \
- a = _mm_packus_epi16 (a, a); \
- pix = _mm_cvtsi128_si32 (a); \
- } while (0)
-
- while ((width -= 4) >= 0)
- {
- INTERPOLATE_ONE_PIXEL (pix1);
- INTERPOLATE_ONE_PIXEL (pix2);
- INTERPOLATE_ONE_PIXEL (pix3);
- INTERPOLATE_ONE_PIXEL (pix4);
- *out++ = pix1;
- *out++ = pix2;
- *out++ = pix3;
- *out++ = pix4;
- }
- if (width & 2)
- {
- INTERPOLATE_ONE_PIXEL (pix1);
- INTERPOLATE_ONE_PIXEL (pix2);
- *out++ = pix1;
- *out++ = pix2;
- }
- if (width & 1)
- {
- INTERPOLATE_ONE_PIXEL (pix1);
- *out = pix1;
- }
-
- #undef INTERPOLATE_ONE_PIXEL
-}
+#define BILINEAR_DECLARE_VARIABLES \
+ const __m128i xmm_wt = _mm_set_epi16 (wt, wt, wt, wt, wt, wt, wt, wt); \
+ const __m128i xmm_wb = _mm_set_epi16 (wb, wb, wb, wb, wb, wb, wb, wb); \
+ const __m128i xmm_xorc = _mm_set_epi16 (0, 0, 0, 0, 0xff, 0xff, 0xff, 0xff);\
+ const __m128i xmm_addc = _mm_set_epi16 (0, 0, 0, 0, 1, 1, 1, 1); \
+ const __m128i xmm_ux = _mm_set_epi16 (unit_x, unit_x, unit_x, unit_x, \
+ unit_x, unit_x, unit_x, unit_x); \
+ const __m128i xmm_zero = _mm_setzero_si128 (); \
+ __m128i xmm_x = _mm_set_epi16 (vx, vx, vx, vx, vx, vx, vx, vx)
+
+#define BILINEAR_INTERPOLATE_ONE_PIXEL(pix) \
+do { \
+ __m128i xmm_wh, xmm_lo, xmm_hi, a; \
+ /* fetch 2x2 pixel block into sse2 register */ \
+ uint32_t tl = src_top [pixman_fixed_to_int (vx)]; \
+ uint32_t tr = src_top [pixman_fixed_to_int (vx) + 1]; \
+ uint32_t bl = src_bottom [pixman_fixed_to_int (vx)]; \
+ uint32_t br = src_bottom [pixman_fixed_to_int (vx) + 1]; \
+ a = _mm_set_epi32 (tr, tl, br, bl); \
+ vx += unit_x; \
+ /* vertical interpolation */ \
+ a = _mm_add_epi16 (_mm_mullo_epi16 (_mm_unpackhi_epi8 (a, xmm_zero), \
+ xmm_wt), \
+ _mm_mullo_epi16 (_mm_unpacklo_epi8 (a, xmm_zero), \
+ xmm_wb)); \
+ /* calculate horizontal weights */ \
+ xmm_wh = _mm_add_epi16 (xmm_addc, \
+ _mm_xor_si128 (xmm_xorc, \
+ _mm_srli_epi16 (xmm_x, 8))); \
+ xmm_x = _mm_add_epi16 (xmm_x, xmm_ux); \
+ /* horizontal interpolation */ \
+ xmm_lo = _mm_mullo_epi16 (a, xmm_wh); \
+ xmm_hi = _mm_mulhi_epu16 (a, xmm_wh); \
+ a = _mm_add_epi32 (_mm_unpacklo_epi16 (xmm_lo, xmm_hi), \
+ _mm_unpackhi_epi16 (xmm_lo, xmm_hi)); \
+ /* shift and pack the result */ \
+ a = _mm_srli_epi32 (a, 16); \
+ a = _mm_packs_epi32 (a, a); \
+ a = _mm_packus_epi16 (a, a); \
+ pix = _mm_cvtsi128_si32 (a); \
+} while (0)
+
+#define BILINEAR_SKIP_ONE_PIXEL() \
+do { \
+ vx += unit_x; \
+ xmm_x = _mm_add_epi16 (xmm_x, xmm_ux); \
+} while(0)
static force_inline void
scaled_bilinear_scanline_sse2_8888_8888_SRC (uint32_t * dst,
pixman_fixed_t max_vx,
pixman_bool_t zero_src)
{
- bilinear_interpolate_line_sse2 (dst, src_top, src_bottom,
- wt, wb, vx, unit_x, w);
+ BILINEAR_DECLARE_VARIABLES;
+ uint32_t pix1, pix2, pix3, pix4;
+
+ while ((w -= 4) >= 0)
+ {
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix2);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix3);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix4);
+ *dst++ = pix1;
+ *dst++ = pix2;
+ *dst++ = pix3;
+ *dst++ = pix4;
+ }
+
+ if (w & 2)
+ {
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix2);
+ *dst++ = pix1;
+ *dst++ = pix2;
+ }
+
+ if (w & 1)
+ {
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+ *dst = pix1;
+ }
+
}
FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8888_cover_SRC,
scaled_bilinear_scanline_sse2_8888_8888_SRC,
uint32_t, uint32_t, uint32_t,
- COVER, FALSE, FALSE)
+ COVER, FLAG_NONE)
FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8888_pad_SRC,
scaled_bilinear_scanline_sse2_8888_8888_SRC,
uint32_t, uint32_t, uint32_t,
- PAD, FALSE, FALSE)
+ PAD, FLAG_NONE)
FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8888_none_SRC,
scaled_bilinear_scanline_sse2_8888_8888_SRC,
uint32_t, uint32_t, uint32_t,
- NONE, FALSE, FALSE)
+ NONE, FLAG_NONE)
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8888_normal_SRC,
+ scaled_bilinear_scanline_sse2_8888_8888_SRC,
+ uint32_t, uint32_t, uint32_t,
+ NORMAL, FLAG_NONE)
+
+static force_inline void
+scaled_bilinear_scanline_sse2_8888_8888_OVER (uint32_t * dst,
+ const uint32_t * mask,
+ const uint32_t * src_top,
+ const uint32_t * src_bottom,
+ int32_t w,
+ int wt,
+ int wb,
+ pixman_fixed_t vx,
+ pixman_fixed_t unit_x,
+ pixman_fixed_t max_vx,
+ pixman_bool_t zero_src)
+{
+ BILINEAR_DECLARE_VARIABLES;
+ uint32_t pix1, pix2, pix3, pix4;
+
+ while (w && ((unsigned long)dst & 15))
+ {
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+
+ if (pix1)
+ {
+ pix2 = *dst;
+ *dst = core_combine_over_u_pixel_sse2 (pix1, pix2);
+ }
+
+ w--;
+ dst++;
+ }
+
+ while (w >= 4)
+ {
+ __m128i xmm_src;
+ __m128i xmm_src_hi, xmm_src_lo, xmm_dst_hi, xmm_dst_lo;
+ __m128i xmm_alpha_hi, xmm_alpha_lo;
+
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix2);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix3);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix4);
+
+ xmm_src = _mm_set_epi32 (pix4, pix3, pix2, pix1);
+
+ if (!is_zero (xmm_src))
+ {
+ if (is_opaque (xmm_src))
+ {
+ save_128_aligned ((__m128i *)dst, xmm_src);
+ }
+ else
+ {
+ __m128i xmm_dst = load_128_aligned ((__m128i *)dst);
+
+ unpack_128_2x128 (xmm_src, &xmm_src_lo, &xmm_src_hi);
+ unpack_128_2x128 (xmm_dst, &xmm_dst_lo, &xmm_dst_hi);
+
+ expand_alpha_2x128 (xmm_src_lo, xmm_src_hi, &xmm_alpha_lo, &xmm_alpha_hi);
+ over_2x128 (&xmm_src_lo, &xmm_src_hi, &xmm_alpha_lo, &xmm_alpha_hi,
+ &xmm_dst_lo, &xmm_dst_hi);
+
+ save_128_aligned ((__m128i *)dst, pack_2x128_128 (xmm_dst_lo, xmm_dst_hi));
+ }
+ }
+
+ w -= 4;
+ dst += 4;
+ }
+
+ while (w)
+ {
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+
+ if (pix1)
+ {
+ pix2 = *dst;
+ *dst = core_combine_over_u_pixel_sse2 (pix1, pix2);
+ }
+
+ w--;
+ dst++;
+ }
+}
+
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8888_cover_OVER,
+ scaled_bilinear_scanline_sse2_8888_8888_OVER,
+ uint32_t, uint32_t, uint32_t,
+ COVER, FLAG_NONE)
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8888_pad_OVER,
+ scaled_bilinear_scanline_sse2_8888_8888_OVER,
+ uint32_t, uint32_t, uint32_t,
+ PAD, FLAG_NONE)
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8888_none_OVER,
+ scaled_bilinear_scanline_sse2_8888_8888_OVER,
+ uint32_t, uint32_t, uint32_t,
+ NONE, FLAG_NONE)
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8888_normal_OVER,
+ scaled_bilinear_scanline_sse2_8888_8888_OVER,
+ uint32_t, uint32_t, uint32_t,
+ NORMAL, FLAG_NONE)
+
+static force_inline void
+scaled_bilinear_scanline_sse2_8888_8_8888_OVER (uint32_t * dst,
+ const uint8_t * mask,
+ const uint32_t * src_top,
+ const uint32_t * src_bottom,
+ int32_t w,
+ int wt,
+ int wb,
+ pixman_fixed_t vx,
+ pixman_fixed_t unit_x,
+ pixman_fixed_t max_vx,
+ pixman_bool_t zero_src)
+{
+ BILINEAR_DECLARE_VARIABLES;
+ uint32_t pix1, pix2, pix3, pix4;
+ uint32_t m;
+
+ while (w && ((unsigned long)dst & 15))
+ {
+ uint32_t sa;
+
+ m = (uint32_t) *mask++;
+
+ if (m)
+ {
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+ sa = pix1 >> 24;
+
+ if (sa == 0xff && m == 0xff)
+ {
+ *dst = pix1;
+ }
+ else
+ {
+ __m128i ms, md, ma, msa;
+
+ pix2 = *dst;
+ ma = expand_alpha_rev_1x128 (load_32_1x128 (m));
+ ms = unpack_32_1x128 (pix1);
+ md = unpack_32_1x128 (pix2);
+
+ msa = expand_alpha_rev_1x128 (load_32_1x128 (sa));
+
+ *dst = pack_1x128_32 (in_over_1x128 (&ms, &msa, &ma, &md));
+ }
+ }
+ else
+ {
+ BILINEAR_SKIP_ONE_PIXEL ();
+ }
+
+ w--;
+ dst++;
+ }
+
+ while (w >= 4)
+ {
+ __m128i xmm_src, xmm_src_lo, xmm_src_hi, xmm_srca_lo, xmm_srca_hi;
+ __m128i xmm_dst, xmm_dst_lo, xmm_dst_hi;
+ __m128i xmm_mask, xmm_mask_lo, xmm_mask_hi;
+
+ m = *(uint32_t*)mask;
+
+ if (m)
+ {
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix2);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix3);
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix4);
+
+ xmm_src = _mm_set_epi32 (pix4, pix3, pix2, pix1);
+
+ if (m == 0xffffffff && is_opaque (xmm_src))
+ {
+ save_128_aligned ((__m128i *)dst, xmm_src);
+ }
+ else
+ {
+ xmm_dst = load_128_aligned ((__m128i *)dst);
+
+ xmm_mask = _mm_unpacklo_epi16 (unpack_32_1x128 (m), _mm_setzero_si128());
+
+ unpack_128_2x128 (xmm_src, &xmm_src_lo, &xmm_src_hi);
+ unpack_128_2x128 (xmm_mask, &xmm_mask_lo, &xmm_mask_hi);
+ unpack_128_2x128 (xmm_dst, &xmm_dst_lo, &xmm_dst_hi);
+
+ expand_alpha_2x128 (xmm_src_lo, xmm_src_hi, &xmm_srca_lo, &xmm_srca_hi);
+ expand_alpha_rev_2x128 (xmm_mask_lo, xmm_mask_hi, &xmm_mask_lo, &xmm_mask_hi);
+
+ in_over_2x128 (&xmm_src_lo, &xmm_src_hi, &xmm_srca_lo, &xmm_srca_hi,
+ &xmm_mask_lo, &xmm_mask_hi, &xmm_dst_lo, &xmm_dst_hi);
+
+ save_128_aligned ((__m128i*)dst, pack_2x128_128 (xmm_dst_lo, xmm_dst_hi));
+ }
+ }
+ else
+ {
+ BILINEAR_SKIP_ONE_PIXEL ();
+ BILINEAR_SKIP_ONE_PIXEL ();
+ BILINEAR_SKIP_ONE_PIXEL ();
+ BILINEAR_SKIP_ONE_PIXEL ();
+ }
+
+ w -= 4;
+ dst += 4;
+ mask += 4;
+ }
+
+ while (w)
+ {
+ uint32_t sa;
+
+ m = (uint32_t) *mask++;
+
+ if (m)
+ {
+ BILINEAR_INTERPOLATE_ONE_PIXEL (pix1);
+ sa = pix1 >> 24;
+
+ if (sa == 0xff && m == 0xff)
+ {
+ *dst = pix1;
+ }
+ else
+ {
+ __m128i ms, md, ma, msa;
+
+ pix2 = *dst;
+ ma = expand_alpha_rev_1x128 (load_32_1x128 (m));
+ ms = unpack_32_1x128 (pix1);
+ md = unpack_32_1x128 (pix2);
+
+ msa = expand_alpha_rev_1x128 (load_32_1x128 (sa));
+
+ *dst = pack_1x128_32 (in_over_1x128 (&ms, &msa, &ma, &md));
+ }
+ }
+ else
+ {
+ BILINEAR_SKIP_ONE_PIXEL ();
+ }
+
+ w--;
+ dst++;
+ }
+}
+
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8_8888_cover_OVER,
+ scaled_bilinear_scanline_sse2_8888_8_8888_OVER,
+ uint32_t, uint8_t, uint32_t,
+ COVER, FLAG_HAVE_NON_SOLID_MASK)
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8_8888_pad_OVER,
+ scaled_bilinear_scanline_sse2_8888_8_8888_OVER,
+ uint32_t, uint8_t, uint32_t,
+ PAD, FLAG_HAVE_NON_SOLID_MASK)
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8_8888_none_OVER,
+ scaled_bilinear_scanline_sse2_8888_8_8888_OVER,
+ uint32_t, uint8_t, uint32_t,
+ NONE, FLAG_HAVE_NON_SOLID_MASK)
+FAST_BILINEAR_MAINLOOP_COMMON (sse2_8888_8_8888_normal_OVER,
+ scaled_bilinear_scanline_sse2_8888_8_8888_OVER,
+ uint32_t, uint8_t, uint32_t,
+ NORMAL, FLAG_HAVE_NON_SOLID_MASK)
static const pixman_fast_path_t sse2_fast_paths[] =
{
SIMPLE_BILINEAR_FAST_PATH (SRC, a8r8g8b8, x8r8g8b8, sse2_8888_8888),
SIMPLE_BILINEAR_FAST_PATH (SRC, x8r8g8b8, x8r8g8b8, sse2_8888_8888),
+ SIMPLE_BILINEAR_FAST_PATH (OVER, a8r8g8b8, x8r8g8b8, sse2_8888_8888),
+ SIMPLE_BILINEAR_FAST_PATH (OVER, a8b8g8r8, x8b8g8r8, sse2_8888_8888),
+ SIMPLE_BILINEAR_FAST_PATH (OVER, a8r8g8b8, a8r8g8b8, sse2_8888_8888),
+ SIMPLE_BILINEAR_FAST_PATH (OVER, a8b8g8r8, a8b8g8r8, sse2_8888_8888),
+
+ SIMPLE_BILINEAR_A8_MASK_FAST_PATH (OVER, a8r8g8b8, x8r8g8b8, sse2_8888_8_8888),
+ SIMPLE_BILINEAR_A8_MASK_FAST_PATH (OVER, a8b8g8r8, x8b8g8r8, sse2_8888_8_8888),
+ SIMPLE_BILINEAR_A8_MASK_FAST_PATH (OVER, a8r8g8b8, a8r8g8b8, sse2_8888_8_8888),
+ SIMPLE_BILINEAR_A8_MASK_FAST_PATH (OVER, a8b8g8r8, a8b8g8r8, sse2_8888_8_8888),
+
{ PIXMAN_OP_NONE },
};
int height = iter->height;
#define FLAGS \
- (FAST_PATH_STANDARD_FLAGS | FAST_PATH_ID_TRANSFORM)
+ (FAST_PATH_STANDARD_FLAGS | FAST_PATH_ID_TRANSFORM | FAST_PATH_BITS_IMAGE)
if ((iter->flags & ITER_NARROW) &&
(image->common.flags & FLAGS) == FLAGS &&
projects / profile / ivi / pixman.git / blobdiff