cheng/libpng
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Prefer memcpy to manual pointer arithmetic.

Browse Source
This commit is contained in:
Samuel Williams 2017-11-09 16:05:00 +13:00 committed by Cosmin Truta
parent a294c1bcb6
commit d3d1baadb4
1 changed files with 10 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
intel/filter_sse2_intrinsics.c
View File

@ -29,39 +29,25 @@
*/ */
static __m128i load4(const void* p) { static __m128i load4(const void* p) {
return _mm_cvtsi32_si128(*(const int*)p); int tmp;
memcpy(&tmp, p, sizeof(tmp));
return _mm_cvtsi32_si128(tmp);
} }
static void store4(void* p, __m128i v) { static void store4(void* p, __m128i v) {
*(int*)p = _mm_cvtsi128_si32(v); int tmp = _mm_cvtsi128_si32(v);
memcpy(p, &tmp, sizeof(int));
} }
static __m128i load3(const void* p) { static __m128i load3(const void* p) {
/* We'll load 2 bytes, then 1 byte, png_uint_32 tmp = 0;
* then mask them together, and finally load into SSE. memcpy(&tmp, p, 3);
*/ return _mm_cvtsi32_si128(tmp);
const png_uint_16* p01 = (png_const_uint_16p)p;
const png_byte* p2 = (const png_byte*)(p01+1);
png_uint_32 v012 = (png_uint_32)(*p01)
| (png_uint_32)(*p2) << 16;
return load4(&v012);
} }
static void store3(void* p, __m128i v) { static void store3(void* p, __m128i v) {
/* We'll pull from SSE as a 32-bit int, then write int tmp = _mm_cvtsi128_si32(v);
* its bottom two bytes, then its third byte. memcpy(p, &tmp, 3);
*/
png_uint_32 v012;
png_uint_16* p01;
png_byte* p2;
store4(&v012, v);
p01 = (png_uint_16p)p;
p2 = (png_byte*)(p01+1);
*p01 = (png_uint_16)v012;
*p2 = (png_byte)(v012 >> 16);
} }
void png_read_filter_row_sub3_sse2(png_row_infop row_info, png_bytep row, void png_read_filter_row_sub3_sse2(png_row_infop row_info, png_bytep row,