X-Git-Url: http://git.osdn.net/view?a=blobdiff_plain;f=source%2Fsrc%2Fcommon.h;h=624eb68dc65bcedae59af15d2c254e0a7703e434;hb=825bf2ca6d1fdd3b866e72eaf7e6db32a4a704c4;hp=7ededd5f503994033f63f3349cc008032c14533d;hpb=7d0e43eeee51ec42e34468f2e9cdc4230e1b2baa;p=csp-qt%2Fcommon_source_project-fm7.git diff --git a/source/src/common.h b/source/src/common.h index 7ededd5f5..624eb68dc 100644 --- a/source/src/common.h +++ b/source/src/common.h @@ -11,17 +11,22 @@ #define _COMMON_H_ // move shared codes to DLL??? +//#ifdef _USE_QT +// #define USE_SHARED_DLL +//#endif + #ifdef _USE_QT - #define USE_SHARED_DLL +#include #endif // use zlib to decompress gzip file??? #ifdef _WIN32 #if defined(_MSC_VER) && (_MSC_VER >= 1500) - #define USE_ZLIB + #ifndef _ANY2D88 + #define USE_ZLIB + #endif #endif #endif - // check environemnt/language #ifdef _WIN32 #ifdef _MSC_VER @@ -45,6 +50,7 @@ // Microsoft Visual C++ 12.0 (2013) or later #define SUPPORT_CPLUSPLUS_11 #endif + #define CSP_OS_WINDOWS #else // Win32, but not Microsoft Visual C++ #define SUPPORT_TCHAR_TYPE @@ -105,6 +111,11 @@ #include #include #include +#ifdef _MSC_VER +#include +#else +#include +#endif #include #include @@ -260,6 +271,90 @@ typedef union { struct { #ifdef __BIG_ENDIAN__ + uint8_t h, l; +#else + uint8_t l, h; +#endif + } b; + struct { +#ifdef __BIG_ENDIAN__ + int8_t h, l; +#else + int8_t l, h; +#endif + } sb; + uint16_t u16; // ToDo: Remove + int16_t s16; // ToDo: Remove + uint16_t w; + int16_t sw; + + inline void read_2bytes_le_from(uint8_t *t) + { + b.l = t[0]; b.h = t[1]; + } + inline void write_2bytes_le_to(uint8_t *t) + { + t[0] = b.l; t[1] = b.h; + } + inline void read_2bytes_be_from(uint8_t *t) + { + b.h = t[0]; b.l = t[1]; + } + inline void write_2bytes_be_to(uint8_t *t) + { + t[0] = b.h; t[1] = b.l; + } + + inline void set_2bytes_be_from(uint16_t n) + { + union { + uint16_t w; + struct { + uint8_t h, l; + }b; + } bigv; + bigv.w = n; + b.l = bigv.b.l; b.h = bigv.b.h; + } + inline void set_2bytes_le_from(uint16_t n) + { + union { + uint16_t w; + struct { + uint8_t l, h; + }b; + } littlev; + littlev.w = n; + b.l = littlev.b.l; b.h = littlev.b.h; + } + inline uint16_t get_2bytes_be_to() + { + union { + uint16_t w; + struct { + uint8_t h, l; + }b; + } bigv; + bigv.b.l = b.l; bigv.b.h = b.h; + return bigv.w; + } + inline uint16_t get_2bytes_le_to() + { + union { + uint16_t w; + struct { + uint8_t l, h; + }b; + } littlev; + littlev.b.l = b.l; littlev.b.h = b.h; + return littlev.w; + } + +} pair16_t; + +typedef union { + struct { +#ifdef __BIG_ENDIAN__ uint8_t h3, h2, h, l; #else uint8_t l, h, h2, h3; @@ -286,8 +381,17 @@ typedef union { int16_t l, h; #endif } sw; + struct { +#ifdef __BIG_ENDIAN__ + pair16_t h, l; +#else + pair16_t l, h; +#endif + } p16; uint32_t d; int32_t sd; + float f; // single float + inline void read_2bytes_le_from(uint8_t *t) { b.l = t[0]; b.h = t[1]; b.h2 = b.h3 = 0; @@ -320,18 +424,393 @@ typedef union { { t[0] = b.h3; t[1] = b.h2; t[2] = b.h; t[3] = b.l; } -} pair_t; + + inline void set_2bytes_be_from(uint16_t n) + { + union { + uint16_t w; + struct { + uint8_t h, l; + }b; + } bigv; + bigv.w = n; + b.l = bigv.b.l; b.h = bigv.b.h; + b.h2 = 0; b.h3 = 0; + } + inline void set_2bytes_le_from(uint16_t n) + { + union { + uint16_t w; + struct { + uint8_t l, h; + }b; + } littlev; + littlev.w = n; + b.l = littlev.b.l; b.h = littlev.b.h; + b.h2 = 0; b.h3 = 0; + } + inline uint16_t get_2bytes_be_to() + { + union { + uint16_t w; + struct { + uint8_t h, l; + }b; + } bigv; + bigv.b.l = b.l; bigv.b.h = b.h; + return bigv.w; + } + inline uint16_t get_2bytes_le_to() + { + union { + uint16_t w; + struct { + uint8_t l, h; + }b; + } littlev; + littlev.b.l = b.l; littlev.b.h = b.h; + return littlev.w; + } + + inline void set_4bytes_be_from(uint32_t n) + { + union { + uint32_t dw; + struct { + uint8_t h3, h2, h, l; + }b; + } bigv; + bigv.dw = n; + b.l = bigv.b.l; b.h = bigv.b.h; b.h2 = bigv.b.h2; b.h3 = bigv.b.h3; + } + inline void set_4bytes_le_from(uint32_t n) + { + union { + uint32_t dw; + struct { + uint8_t l, h, h2, h3; + }b; + } littlev; + littlev.dw = n; + b.l = littlev.b.l; b.h = littlev.b.h; b.h2 = littlev.b.h2; b.h3 = littlev.b.h3; + } + inline uint32_t get_4bytes_be_to() + { + union { + uint32_t dw; + struct { + uint8_t h3, h2, h, l; + }b; + } bigv; + bigv.b.l = b.l; bigv.b.h = b.h; bigv.b.h2 = b.h2; bigv.b.h3 = b.h3; + return bigv.dw; + } + inline uint32_t get_4bytes_le_to() + { + union { + uint32_t dw; + struct { + uint8_t l, h, h2, h3; + }b; + } littlev; + littlev.b.l = b.l; littlev.b.h = b.h; littlev.b.h2 = b.h2; littlev.b.h3 = b.h3; + return littlev.dw; + } +} pair32_t; + + +typedef union { + struct { +#ifdef __BIG_ENDIAN__ + uint8_t h7, h6, h5, h4, h3, h2, h, l; +#else + uint8_t l, h, h2, h3, h4, h5, h6, h7; +#endif + } b; + struct { +#ifdef __BIG_ENDIAN__ + int8_t h7, h6, h5, h4, h3, h2, h, l; +#else + int8_t l, h, h2, h3, h4, h5, h6, h7; +#endif + } sb; + struct { +#ifdef __BIG_ENDIAN__ + uint16_t h3, h2, h, l; +#else + uint16_t l, h, h2, h3; +#endif + } w; + struct { +#ifdef __BIG_ENDIAN__ + int16_t h3, h2, h, l; +#else + int16_t l, h, h2, h3; +#endif + } sw; + struct { +#ifdef __BIG_ENDIAN__ + pair16_t h3, h2, h, l; +#else + pair16_t l, h, h2, h3; +#endif + } p16; + struct { +#ifdef __BIG_ENDIAN__ + uint32_t h, l; +#else + uint32_t l, h; +#endif + } d; + struct { +#ifdef __BIG_ENDIAN__ + int32_t h, l; +#else + int32_t l, h; +#endif + } sd; + struct { +#ifdef __BIG_ENDIAN__ + pair32_t h, l; +#else + pair32_t l, h; +#endif + } p32; + struct { +#ifdef __BIG_ENDIAN__ + float h, l; +#else + float l, h; +#endif + } f32; + uint64_t q; + int64_t sq; + double df; // double float + inline void read_2bytes_le_from(uint8_t *t) + { + b.l = t[0]; b.h = t[1]; b.h2 = b.h3 = 0; + b.h4 = 0; b.h5 = 0; b.h6 = 0; b.h7 = 0; + } + inline void write_2bytes_le_to(uint8_t *t) + { + t[0] = b.l; t[1] = b.h; + } + inline void read_2bytes_be_from(uint8_t *t) + { + b.h3 = b.h2 = 0; b.h = t[0]; b.l = t[1]; + b.h4 = 0; b.h5 = 0; b.h6 = 0; b.h7 = 0; + } + inline void write_2bytes_be_to(uint8_t *t) + { + t[0] = b.h; t[1] = b.l; + } + inline void read_4bytes_le_from(uint8_t *t) + { + b.l = t[0]; b.h = t[1]; b.h2 = t[2]; b.h3 = t[3]; + b.h4 = 0; b.h5 = 0; b.h6 = 0; b.h7 = 0; + } + inline void write_4bytes_le_to(uint8_t *t) + { + t[0] = b.l; t[1] = b.h; t[2] = b.h2; t[3] = b.h3; + } + inline void read_4bytes_be_from(uint8_t *t) + { + b.h3 = t[0]; b.h2 = t[1]; b.h = t[2]; b.l = t[3]; + b.h4 = 0; b.h5 = 0; b.h6 = 0; b.h7 = 0; + } + inline void write_4bytes_be_to(uint8_t *t) + { + t[0] = b.h3; t[1] = b.h2; t[2] = b.h; t[3] = b.l; + } + + inline void read_8bytes_le_from(uint8_t *t) + { + b.l = t[0]; b.h = t[1]; b.h2 = t[2]; b.h3 = t[3]; + b.h4 = t[4]; b.h5 = t[5]; b.h6 = t[6]; b.h7 = t[7]; + } + inline void write_8bytes_le_to(uint8_t *t) + { + t[0] = b.l; t[1] = b.h; t[2] = b.h2; t[3] = b.h3; + t[4] = b.h4; t[5] = b.h5; t[6] = b.h6; t[7] = b.h7; + } + inline void read_8bytes_be_from(uint8_t *t) + { + b.h7 = t[0]; b.h6 = t[1]; b.h5 = t[2]; b.h4 = t[3]; + b.h3 = t[4]; b.h2 = t[5]; b.h = t[6]; b.l = t[7]; + } + inline void write_8bytes_be_to(uint8_t *t) + { + t[0] = b.h7; t[1] = b.h6; t[2] = b.h5; t[3] = b.h4; + t[4] = b.h3; t[5] = b.h2; t[6] = b.h; t[7] = b.l; + } + + inline void set_2bytes_be_from(uint16_t n) + { + union { + uint16_t w; + struct { + uint8_t h, l; + }b; + } bigv; + bigv.w = n; + b.l = bigv.b.l; b.h = bigv.b.h; + b.h2 = 0; b.h3 = 0; + b.h4 = 0; b.h5 = 0; b.h6 = 0; b.h7 = 0; + } + inline void set_2bytes_le_from(uint16_t n) + { + union { + uint16_t w; + struct { + uint8_t l, h; + }b; + } littlev; + littlev.w = n; + b.l = littlev.b.l; b.h = littlev.b.h; + b.h2 = 0; b.h3 = 0; + b.h4 = 0; b.h5 = 0; b.h6 = 0; b.h7 = 0; + } + inline uint16_t get_2bytes_be_to() + { + union { + uint16_t w; + struct { + uint8_t h, l; + }b; + } bigv; + bigv.b.l = b.l; bigv.b.h = b.h; + return bigv.w; + } + inline uint16_t get_2bytes_le_to() + { + union { + uint16_t w; + struct { + uint8_t l, h; + }b; + } littlev; + littlev.b.l = b.l; littlev.b.h = b.h; + return littlev.w; + } + + inline void set_4bytes_be_from(uint32_t n) + { + union { + uint32_t dw; + struct { + uint8_t h3, h2, h, l; + }b; + } bigv; + bigv.dw = n; + b.l = bigv.b.l; b.h = bigv.b.h; b.h2 = bigv.b.h2; b.h3 = bigv.b.h3; + b.h4 = 0; b.h5 = 0; b.h6 = 0; b.h7 = 0; + } + inline void set_4bytes_le_from(uint32_t n) + { + union { + uint32_t dw; + struct { + uint8_t l, h, h2, h3; + }b; + } littlev; + littlev.dw = n; + b.l = littlev.b.l; b.h = littlev.b.h; b.h2 = littlev.b.h2; b.h3 = littlev.b.h3; + b.h4 = 0; b.h5 = 0; b.h6 = 0; b.h7 = 0; + } + inline uint32_t get_4bytes_be_to() + { + union { + uint32_t dw; + struct { + uint8_t h3, h2, h, l; + }b; + } bigv; + bigv.b.l = b.l; bigv.b.h = b.h; bigv.b.h2 = b.h2; bigv.b.h3 = b.h3; + return bigv.dw; + } + inline uint32_t get_4bytes_le_to() + { + union { + uint32_t dw; + struct { + uint8_t l, h, h2, h3; + }b; + } littlev; + littlev.b.l = b.l; littlev.b.h = b.h; littlev.b.h2 = b.h2; littlev.b.h3 = b.h3; + return littlev.dw; + } + + inline void set_8bytes_be_from(uint64_t n) + { + union { + uint64_t qw; + struct { + uint8_t h7, h6, h5, h4, h3, h2, h, l; + }b; + } bigv; + bigv.qw = n; + b.l = bigv.b.l; b.h = bigv.b.h; b.h2 = bigv.b.h2; b.h3 = bigv.b.h3; + b.h4 = bigv.b.h4; b.h5 = bigv.b.h5; b.h6 = bigv.b.h6; b.h7 = bigv.b.h7; + } + inline void set_8bytes_le_from(uint64_t n) + { + union { + uint64_t qw; + struct { + uint8_t l, h, h2, h3, h4, h5, h6, h7; + }b; + } littlev; + littlev.qw = n; + b.l = littlev.b.l; b.h = littlev.b.h; b.h2 = littlev.b.h2; b.h3 = littlev.b.h3; + b.h4 = littlev.b.h4; b.h5 = littlev.b.h5; b.h6 = littlev.b.h6; b.h7 = littlev.b.h7; + } + inline uint64_t get_8bytes_be_to() + { + union { + uint64_t qw; + struct { + uint8_t h7, h6, h5, h4, h3, h2, h, l; + }b; + } bigv; + bigv.b.l = b.l; bigv.b.h = b.h; bigv.b.h2 = b.h2; bigv.b.h3 = b.h3; + bigv.b.h4 = b.h4; bigv.b.h5 = b.h5; bigv.b.h6 = b.h6; bigv.b.h7 = b.h7; + return bigv.qw; + } + inline uint64_t get_8bytes_le_to() + { + union { + uint64_t qw; + struct { + uint8_t l, h, h2, h3, h4, h5, h6, h7; + }b; + } littlev; + littlev.b.l = b.l; littlev.b.h = b.h; littlev.b.h2 = b.h2; littlev.b.h3 = b.h3; + littlev.b.h4 = b.h4; littlev.b.h5 = b.h5; littlev.b.h6 = b.h6; littlev.b.h7 = b.h7; + return littlev.qw; + } + +} pair64_t; uint32_t DLL_PREFIX EndianToLittle_DWORD(uint32_t x); uint16_t DLL_PREFIX EndianToLittle_WORD(uint16_t x); +uint32_t DLL_PREFIX EndianFromLittle_DWORD(uint32_t x); +uint16_t DLL_PREFIX EndianFromLittle_WORD(uint16_t x); +uint32_t DLL_PREFIX EndianToBig_DWORD(uint32_t x); +uint16_t DLL_PREFIX EndianToBig_WORD(uint16_t x); +uint32_t DLL_PREFIX EndianFromBig_DWORD(uint32_t x); +uint16_t DLL_PREFIX EndianFromBig_WORD(uint16_t x); // max/min #ifndef _MSC_VER #undef max #undef min int DLL_PREFIX max(int a, int b); + unsigned int DLL_PREFIX max(int a, unsigned int b); + unsigned int DLL_PREFIX max(unsigned int a, int b); unsigned int DLL_PREFIX max(unsigned int a, unsigned int b); int DLL_PREFIX min(int a, int b); + int DLL_PREFIX min(unsigned int a, int b); + int DLL_PREFIX min(int a, unsigned int b); unsigned int DLL_PREFIX min(unsigned int a, unsigned int b); #endif @@ -427,7 +906,7 @@ uint16_t DLL_PREFIX EndianToLittle_WORD(uint16_t x); errno_t DLL_PREFIX my_strncpy_s(char *strDestination, size_t numberOfElements, const char *strSource, size_t count); errno_t DLL_PREFIX my_tcsncpy_s(_TCHAR *strDestination, size_t numberOfElements, const _TCHAR *strSource, size_t count); char * DLL_PREFIX my_strtok_s(char *strToken, const char *strDelimit, char **context); - _TCHAR * DLL_PREFIX my_tcstok_s(_TCHAR *strToken, const char *strDelimit, _TCHAR **context); + _TCHAR *DLL_PREFIX my_tcstok_s(_TCHAR *strToken, const char *strDelimit, _TCHAR **context); #define my_fprintf_s fprintf #define my_ftprintf_s fprintf int DLL_PREFIX my_sprintf_s(char *buffer, size_t sizeOfBuffer, const char *format, ...); @@ -453,12 +932,29 @@ uint16_t DLL_PREFIX EndianToLittle_WORD(uint16_t x); #define my_vstprintf_s _vstprintf_s #endif +// memory +#ifndef _MSC_VER + void *DLL_PREFIX my_memcpy(void *dst, void *src, size_t len); +#else + #define my_memcpy memcpy +#endif + +// hint for SIMD +#if defined(__clang__) + #define __DECL_VECTORIZED_LOOP _Pragma("clang loop vectorize(enable) distribute(enable)") +#elif defined(__GNUC__) + #define __DECL_VECTORIZED_LOOP _Pragma("GCC ivdep") +#else + #define __DECL_VECTORIZED_LOOP +#endif + // C99 math functions #ifdef _MSC_VER - #define my_isfinite _finite + #define my_isfinite _finite #define my_log2(v) (log((double)(v)) / log(2.0)) #else - #define my_isfinite isfinite + #include + #define my_isfinite std::isfinite #define my_log2 log2 #endif @@ -491,13 +987,365 @@ uint16_t DLL_PREFIX EndianToLittle_WORD(uint16_t x); uint8_t DLL_PREFIX A_OF_COLOR(scrntype_t c); #elif defined(_RGB888) typedef uint32_t scrntype_t; - #define RGB_COLOR(r, g, b) (((uint32_t)(r) << 16) | ((uint32_t)(g) << 8) | ((uint32_t)(b) << 0)) +#if defined(__LITTLE_ENDIAN__) + #define RGB_COLOR(r, g, b) (((uint32_t)(b) << 16) | ((uint32_t)(g) << 8) | ((uint32_t)(r) << 0) | (0xff << 24)) + #define RGBA_COLOR(r, g, b, a) (((uint32_t)(b) << 16) | ((uint32_t)(g) << 8) | ((uint32_t)(r) << 0) | ((uint32_t)(a) << 24)) + #define R_OF_COLOR(c) (((c) ) & 0xff) + #define G_OF_COLOR(c) (((c) >> 8) & 0xff) + #define B_OF_COLOR(c) (((c) >> 16) & 0xff) + #define A_OF_COLOR(c) (((c) >> 24) & 0xff) +#else + #define RGB_COLOR(r, g, b) (((uint32_t)(r) << 16) | ((uint32_t)(g) << 8) | ((uint32_t)(b) << 0) | (0xff << 24)) #define RGBA_COLOR(r, g, b, a) (((uint32_t)(r) << 16) | ((uint32_t)(g) << 8) | ((uint32_t)(b) << 0) | ((uint32_t)(a) << 24)) #define R_OF_COLOR(c) (((c) >> 16) & 0xff) #define G_OF_COLOR(c) (((c) >> 8) & 0xff) #define B_OF_COLOR(c) (((c) ) & 0xff) #define A_OF_COLOR(c) (((c) >> 24) & 0xff) #endif +#endif + +// 20181104 K.O: +// Below routines aims to render common routine. + +#ifdef _MSC_VER + #define __DECL_ALIGNED(foo) __declspec(align(foo)) + #ifndef __builtin_assume_aligned + #define __builtin_assume_aligned(foo, a) foo + #endif +#elif defined(__GNUC__) + #define __DECL_ALIGNED(foo) __attribute__((aligned(foo))) +#else + // ToDo + #define __builtin_assume_aligned(foo, a) foo + #define __DECL_ALIGNED(foo) +#endif + +// ToDo: for MSVC +#if defined(_RGB555) || defined(_RGBA565) +typedef __DECL_ALIGNED(16) union { + scrntype_t w[8]; + __v8hi v; +} scrntype_vec8_t; +typedef __DECL_ALIGNED(16) union { + scrntype_t w[16]; + __v8hi v[2]; +} scrntype_vec16_t; +#else +typedef __DECL_ALIGNED(32) union { + scrntype_t w[8]; + __v16hi v; +} scrntype_vec8_t; +typedef __DECL_ALIGNED(32) union { + scrntype_t w[16]; + __v16hi v[2]; +} scrntype_vec16_t; +#endif + +typedef __DECL_ALIGNED(16) union { + __v4hi v; + uint8_t w[8]; +} uint8_vec8_t; + +typedef __DECL_ALIGNED(16) union { + __v8hi v; + uint16_t w[8]; +} uint16_vec8_t; + +typedef __DECL_ALIGNED(16) union { + __v16hi v; + uint32_t w[8]; +} uint32_vec8_t; + +typedef __DECL_ALIGNED(16) struct { + uint16_vec8_t plane_table[256]; +} _bit_trans_table_t; + +typedef __DECL_ALIGNED(sizeof(scrntype_vec8_t)) struct { + scrntype_vec8_t plane_table[256]; +} _bit_trans_table_scrn_t; + +typedef struct { + scrntype_t* palette; // Must be 2^planes entries. If NULL, assume RGB. + _bit_trans_table_t* bit_trans_table[16]; // Must be exist >= planes. Must be aligned with sizeof(uint16_vec8_t). + int xzoom; // 1 - 4? + bool is_render[16]; + int shift; + uint8_t* data[16]; + uint32_t baseaddress[16]; + uint32_t voffset[16]; + uint32_t addrmask; // For global increment. + uint32_t addrmask2; // For local increment. + uint32_t begin_pos; + uint32_t render_width; +} _render_command_data_t; + + +inline scrntype_vec8_t ConvertByteToMonochromePackedPixel(uint8_t src, _bit_trans_table_t *tbl,scrntype_t on_val, scrntype_t off_val) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + __DECL_ALIGNED(32) scrntype_vec8_t tmpdd; + _bit_trans_table_t* vt = (_bit_trans_table_t*)__builtin_assume_aligned(tbl, sizeof(uint16_vec8_t)); + + tmpd.v = vt->plane_table[src].v; +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 8; i++) { + tmpdd.w[i] = (tmpd.w[i] == 0) ? off_val: on_val; + } + return tmpdd; +} + +// Note: Pls. read Note(s) of common.cpp -- 20181105 K.Ohta. +// Tables for below functions must be aligned by 16 (_bit_trans_table_t) or 32(_bit_trans_table_scrn_t). +void DLL_PREFIX ConvertByteToPackedPixelByColorTable(uint8_t *src, scrntype_t* dst, int bytes, _bit_trans_table_t *tbl, scrntype_t *on_color_table, scrntype_t* off_color_table); +void DLL_PREFIX ConvertByteToPackedPixelByColorTable2(uint8_t *src, scrntype_t* dst, int bytes, _bit_trans_table_scrn_t *tbl, scrntype_t *on_color_table, scrntype_t* off_color_table); +void DLL_PREFIX ConvertByteToSparceUint16(uint8_t *src, uint16_t* dst, int bytes, _bit_trans_table_t *tbl, uint16_t mask); +void DLL_PREFIX ConvertByteToSparceUint8(uint8_t *src, uint16_t* dst, int bytes, _bit_trans_table_t *tbl, uint16_t mask); + +// Table must be (ON_VAL_COLOR : OFF_VAL_COLOR)[256]. +inline scrntype_vec8_t ConvertByteToPackedPixel_PixelTbl(uint8_t src, _bit_trans_table_scrn_t *tbl) +{ + __DECL_ALIGNED(32) scrntype_vec8_t tmpdd; + _bit_trans_table_scrn_t* vt = (_bit_trans_table_scrn_t*)__builtin_assume_aligned(tbl, sizeof(uint16_vec8_t)); + + tmpdd.v = vt->plane_table[src].v; + return tmpdd; +} + +// Table must be (ON_VAL_COLOR : OFF_VAL_COLOR)[256]. +inline scrntype_vec16_t ConvertByteToDoublePackedPixel_PixelTbl(uint8_t src, _bit_trans_table_scrn_t *tbl) +{ + __DECL_ALIGNED(32) scrntype_vec16_t tmpdd; + __DECL_ALIGNED(32) scrntype_vec8_t tmpd; + _bit_trans_table_scrn_t* vt = (_bit_trans_table_scrn_t*)__builtin_assume_aligned(tbl, sizeof(uint16_vec8_t)); + tmpd.v = vt->plane_table[src].v; + int j = 0; +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 16; i += 2) { + tmpdd.w[i] = tmpd.w[j]; + tmpdd.w[i + 1] = tmpd.w[j]; + j++; + } + return tmpdd; +} + +// Table must be initialize ON_COLOR : OFF_COLOR +inline void ConvertByteToDoubleMonochromeUint8(uint8_t src, uint8_t* dst, _bit_trans_table_t* tbl) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + uint16_vec8_t* vt = (uint16_vec8_t*)__builtin_assume_aligned(&(tbl->plane_table[0]), sizeof(uint16_vec8_t)); + + __DECL_ALIGNED(16) uint8_t d[16]; + tmpd = vt[src]; + int j = 0; +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 16; i += 2) { + d[i] = (uint8_t)(tmpd.w[j]); + d[i + 1] = (uint8_t)(tmpd.w[j]); + j++; + } +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 16; i++) { + dst[i] = d[i]; + } +} + +inline void ConvertByteToMonochromeUint8(uint8_t src, uint8_t* dst, _bit_trans_table_t* tbl) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + uint16_vec8_t* vt = (uint16_vec8_t*)__builtin_assume_aligned(&(tbl->plane_table[0]), sizeof(uint16_vec8_t)); + + tmpd = vt[src]; +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 8; i++) { + dst[i] = (uint8_t)(tmpd.w[i]); + } +} + +inline void ConvertRGBTo8ColorsUint8(uint8_t r, uint8_t g, uint8_t b, uint8_t* dst, _bit_trans_table_t* rtbl, _bit_trans_table_t* gtbl, _bit_trans_table_t* btbl, int shift) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + uint16_vec8_t* rvt = (uint16_vec8_t*)__builtin_assume_aligned(&(rtbl->plane_table[0]), sizeof(uint16_vec8_t)); + uint16_vec8_t* gvt = (uint16_vec8_t*)__builtin_assume_aligned(&(gtbl->plane_table[0]), sizeof(uint16_vec8_t)); + uint16_vec8_t* bvt = (uint16_vec8_t*)__builtin_assume_aligned(&(btbl->plane_table[0]), sizeof(uint16_vec8_t)); + + tmpd.v = rvt[r].v; + tmpd.v = tmpd.v | gvt[g].v; + tmpd.v = tmpd.v | bvt[b].v; + tmpd.v = tmpd.v >> shift; +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 8; i++) { + dst[i] = (uint8_t)(tmpd.w[i]); + } +} + +inline void ConvertRGBTo8ColorsUint8_Zoom2Left(uint8_t r, uint8_t g, uint8_t b, uint8_t* dst, _bit_trans_table_t* rtbl, _bit_trans_table_t* gtbl, _bit_trans_table_t* btbl, int shift) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + uint16_vec8_t* rvt = (uint16_vec8_t*)__builtin_assume_aligned(&(rtbl->plane_table[0]), sizeof(uint16_vec8_t)); + uint16_vec8_t* gvt = (uint16_vec8_t*)__builtin_assume_aligned(&(gtbl->plane_table[0]), sizeof(uint16_vec8_t)); + uint16_vec8_t* bvt = (uint16_vec8_t*)__builtin_assume_aligned(&(btbl->plane_table[0]), sizeof(uint16_vec8_t)); + + tmpd.v = rvt[r].v; + tmpd.v = tmpd.v | gvt[g].v; + tmpd.v = tmpd.v | bvt[b].v; + tmpd.v = tmpd.v >> shift; +__DECL_VECTORIZED_LOOP + for(int i = 0, j = 0; i < 8; i += 2, j++) { + dst[i] = (uint8_t)(tmpd.w[j]); + dst[i + 1] = (uint8_t)(tmpd.w[j]); + } +} + +inline void ConvertRGBTo8ColorsUint8_Zoom2Right(uint8_t r, uint8_t g, uint8_t b, uint8_t* dst, _bit_trans_table_t* rtbl, _bit_trans_table_t* gtbl, _bit_trans_table_t* btbl, int shift) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + uint16_vec8_t* rvt = (uint16_vec8_t*)__builtin_assume_aligned(&(rtbl->plane_table[0]), sizeof(uint16_vec8_t)); + uint16_vec8_t* gvt = (uint16_vec8_t*)__builtin_assume_aligned(&(gtbl->plane_table[0]), sizeof(uint16_vec8_t)); + uint16_vec8_t* bvt = (uint16_vec8_t*)__builtin_assume_aligned(&(btbl->plane_table[0]), sizeof(uint16_vec8_t)); + + tmpd.v = rvt[r].v; + tmpd.v = tmpd.v | gvt[g].v; + tmpd.v = tmpd.v | bvt[b].v; + tmpd.v = tmpd.v >> shift; +__DECL_VECTORIZED_LOOP + for(int i = 0, j = 4; i < 8; i += 2, j++) { + dst[i] = (uint8_t)(tmpd.w[j]); + dst[i + 1] = (uint8_t)(tmpd.w[j]); + } +} + +inline void ConvertRGBTo8ColorsUint8_Zoom2Double(uint8_t r, uint8_t g, uint8_t b, uint8_t* dst, _bit_trans_table_t* rtbl, _bit_trans_table_t* gtbl, _bit_trans_table_t* btbl, int shift) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + uint16_vec8_t* rvt = (uint16_vec8_t*)__builtin_assume_aligned(&(rtbl->plane_table[0]), sizeof(uint16_vec8_t)); + uint16_vec8_t* gvt = (uint16_vec8_t*)__builtin_assume_aligned(&(gtbl->plane_table[0]), sizeof(uint16_vec8_t)); + uint16_vec8_t* bvt = (uint16_vec8_t*)__builtin_assume_aligned(&(btbl->plane_table[0]), sizeof(uint16_vec8_t)); + + tmpd.v = rvt[r].v; + tmpd.v = tmpd.v | gvt[g].v; + tmpd.v = tmpd.v | bvt[b].v; + tmpd.v = tmpd.v >> shift; +__DECL_VECTORIZED_LOOP + for(int i = 0, j = 0; i < 16; i += 2, j++) { + dst[i] = (uint8_t)(tmpd.w[j]); + dst[i + 1] = (uint8_t)(tmpd.w[j]); + } +} + +inline void ConvertByteToMonochromeUint8Cond_Zoom2(uint8_t src, uint8_t* dst, _bit_trans_table_t* tbl, uint8_t on_color, uint8_t off_color) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + uint16_vec8_t* vt = (uint16_vec8_t*)__builtin_assume_aligned(&(tbl->plane_table[0]), sizeof(uint16_vec8_t)); + + __DECL_ALIGNED(16) uint8_t d[16]; + tmpd = vt[src]; + int j = 0; +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 16; i += 2) { + d[i] = (tmpd.w[j] == 0) ? off_color : on_color; + d[i + 1] = (tmpd.w[j] == 0) ? off_color : on_color; + j++; + } +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 16; i++) { + dst[i] = d[i]; + } +} + +inline void ConvertByteToMonochromeUint8Cond(uint8_t src, uint8_t* dst, _bit_trans_table_t* tbl, uint8_t on_color, uint8_t off_color) +{ + __DECL_ALIGNED(16) uint16_vec8_t tmpd; + uint16_vec8_t* vt = (uint16_vec8_t*)__builtin_assume_aligned(&(tbl->plane_table[0]), sizeof(uint16_vec8_t)); + + tmpd = vt[src]; +__DECL_VECTORIZED_LOOP + for(int i = 0; i < 8; i++) { + dst[i] = (tmpd.w[i] == 0) ? off_color : on_color; + } +} + +void DLL_PREFIX PrepareBitTransTableUint16(_bit_trans_table_t *tbl, uint16_t on_val, uint16_t off_val); +void DLL_PREFIX PrepareBitTransTableScrnType(_bit_trans_table_scrn_t *tbl, scrntype_t on_val, scrntype_t off_val); +void DLL_PREFIX PrepareReverseBitTransTableUint16(_bit_trans_table_t *tbl, uint16_t on_val, uint16_t off_val); +void DLL_PREFIX PrepareReverseBitTransTableScrnType(_bit_trans_table_scrn_t *tbl, scrntype_t on_val, scrntype_t off_val); + +void DLL_PREFIX Render8Colors_Line(_render_command_data_t *src, scrntype_t *dst, scrntype_t *dst2, bool scan_line); + +void DLL_PREFIX Render16Colors_Line(_render_command_data_t *src, scrntype_t *dst, scrntype_t *dst2, bool scan_line); +void DLL_PREFIX Render2NColors_Line(_render_command_data_t *src, scrntype_t *dst, scrntype_t* dst2, bool scan_line, int planes); + +void DLL_PREFIX Convert8ColorsToByte_Line(_render_command_data_t *src, uint8_t *dst); +void DLL_PREFIX Convert2NColorsToByte_Line(_render_command_data_t *src, uint8_t *dst, int planes); +void DLL_PREFIX Convert2NColorsToByte_LineZoom2(_render_command_data_t *src, uint8_t *dst, int planes); + +inline uint64_t ExchangeEndianU64(uint64_t __in) +{ + pair64_t __i, __o; + __i.q = __in; + __o.b.h7 = __i.b.l; + __o.b.h6 = __i.b.h; + __o.b.h5 = __i.b.h2; + __o.b.h4 = __i.b.h3; + __o.b.h3 = __i.b.h4; + __o.b.h2 = __i.b.h5; + __o.b.h = __i.b.h6; + __o.b.l = __i.b.h7; + return __o.q; +} + +inline int64_t ExchangeEndianS64(uint64_t __in) +{ + pair64_t __i, __o; + __i.q = __in; + __o.b.h7 = __i.b.l; + __o.b.h6 = __i.b.h; + __o.b.h5 = __i.b.h2; + __o.b.h4 = __i.b.h3; + __o.b.h3 = __i.b.h4; + __o.b.h2 = __i.b.h5; + __o.b.h = __i.b.h6; + __o.b.l = __i.b.h7; + return __o.sq; +} +inline uint32_t ExchangeEndianU32(uint32_t __in) +{ + pair32_t __i, __o; + __i.d = __in; + __o.b.h3 = __i.b.l; + __o.b.h2 = __i.b.h; + __o.b.h = __i.b.h2; + __o.b.l = __i.b.h3; + return __o.d; +} + +inline int32_t ExchangeEndianS32(uint32_t __in) +{ + pair32_t __i, __o; + __i.d = __in; + __o.b.h3 = __i.b.l; + __o.b.h2 = __i.b.h; + __o.b.h = __i.b.h2; + __o.b.l = __i.b.h3; + return __o.sd; +} + +inline uint16_t ExchangeEndianU16(uint16_t __in) +{ + pair16_t __i, __o; + __i.u16 = __in; + __o.b.h = __i.b.l; + __o.b.l = __i.b.h; + return __o.u16; +} + +inline int16_t ExchangeEndianS16(uint16_t __in) +{ + pair16_t __i, __o; + __i.u16 = __in; + __o.b.h = __i.b.l; + __o.b.l = __i.b.h; + return __o.s16; +} // wav file header #pragma pack(1) @@ -521,12 +1369,41 @@ typedef struct { } wav_header_t; #pragma pack() +// See http://www-mmsp.ece.mcgill.ca/Documents/AudioFormats/WAVE/WAVE.html. +#pragma pack(1) +typedef struct { + wav_chunk_t riff_chunk; + char wave[4]; + wav_chunk_t fmt_chunk; + uint16_t format_id; + uint16_t channels; + uint32_t sample_rate; + uint32_t data_speed; + uint16_t block_size; + uint16_t sample_bits; + uint16_t cbsize; // Extension size.Normaly set to 0. + wav_chunk_t fact_chunk; // "fact", 4. +} wav_header_float_t; +#pragma pack() + +// Use this before writing wav_data. +bool DLL_PREFIX write_dummy_wav_header(void *__fio); +// Use this after writng wav_data. +bool DLL_PREFIX set_wav_header(wav_header_t *header, wav_chunk_t *first_chunk, uint16_t channels, uint32_t rate, + uint16_t bits, size_t file_length); +bool DLL_PREFIX load_wav_to_stereo(void *__fio, int16_t **left_buf, int16_t **right_buf, uint32_t *rate, int *got_samples); +bool DLL_PREFIX load_wav_to_monoral(void *__fio, int16_t **buffer, uint32_t *rate, int *got_samples); + // file path const _TCHAR *DLL_PREFIX get_application_path(); +const _TCHAR *DLL_PREFIX get_initial_current_path(); const _TCHAR *DLL_PREFIX create_local_path(const _TCHAR *format, ...); void DLL_PREFIX create_local_path(_TCHAR *file_path, int length, const _TCHAR *format, ...); const _TCHAR *DLL_PREFIX create_date_file_path(const _TCHAR *extension); +bool DLL_PREFIX is_absolute_path(const _TCHAR *file_path); void DLL_PREFIX create_date_file_path(_TCHAR *file_path, int length, const _TCHAR *extension); +const _TCHAR *DLL_PREFIX create_date_file_name(const _TCHAR *extension); +void DLL_PREFIX create_date_file_name(_TCHAR *file_path, int length, const _TCHAR *extension); bool DLL_PREFIX check_file_extension(const _TCHAR *file_path, const _TCHAR *ext); const _TCHAR *DLL_PREFIX get_file_path_without_extensiton(const _TCHAR *file_path); void DLL_PREFIX get_long_full_path_name(const _TCHAR* src, _TCHAR* dst, size_t dst_len); @@ -541,11 +1418,15 @@ const char *DLL_PREFIX tchar_to_char(const _TCHAR *ts); const _TCHAR *DLL_PREFIX wchar_to_tchar(const wchar_t *ws); const wchar_t *DLL_PREFIX tchar_to_wchar(const _TCHAR *ts); + // misc +void DLL_PREFIX common_initialize(); + int32_t DLL_PREFIX muldiv_s32(int32_t nNumber, int32_t nNumerator, int32_t nDenominator); uint32_t DLL_PREFIX muldiv_u32(uint32_t nNumber, uint32_t nNumerator, uint32_t nDenominator); uint32_t DLL_PREFIX get_crc32(uint8_t data[], int size); +uint32_t DLL_PREFIX calc_crc32(uint32_t seed, uint8_t data[], int size); uint16_t DLL_PREFIX jis_to_sjis(uint16_t jis); int DLL_PREFIX decibel_to_volume(int decibel); @@ -573,11 +1454,11 @@ typedef DLL_PREFIX struct cur_time_s { void increment(); void update_year(); void update_day_of_week(); - void save_state(void *f); - bool load_state(void *f); + bool process_state(void *f, bool loading); } cur_time_t; void DLL_PREFIX get_host_time(cur_time_t* cur_time); +const _TCHAR DLL_PREFIX *get_lib_common_version(); // symbol typedef struct symbol_s {