novidia: add shader.farc xdelta/vcdiff patching

This commit is contained in:
somewhatlurker
2020-11-10 05:00:18 +11:00
parent 56eb3333be
commit 8abf74bf97
98 changed files with 45837 additions and 2 deletions
@@ -57,7 +57,7 @@
<SDLCheck>true</SDLCheck>
<PreprocessorDefinitions>_DEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<ConformanceMode>true</ConformanceMode>
<AdditionalIncludeDirectories>..\..\..\dependencies\PluginConfigApi;..\..\..\dependencies\detours\include;</AdditionalIncludeDirectories>
<AdditionalIncludeDirectories>..\..\..\dependencies\xdelta3;..\..\..\dependencies\PluginConfigApi;..\..\..\dependencies\detours\include;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
@@ -74,7 +74,7 @@
<SDLCheck>true</SDLCheck>
<PreprocessorDefinitions>NDEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<ConformanceMode>true</ConformanceMode>
<AdditionalIncludeDirectories>..\..\..\dependencies\PluginConfigApi;..\..\..\dependencies\detours\include;</AdditionalIncludeDirectories>
<AdditionalIncludeDirectories>..\..\..\dependencies\xdelta3;..\..\..\dependencies\PluginConfigApi;..\..\..\dependencies\detours\include;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
@@ -86,9 +86,12 @@
</Link>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="src\crc\crc_32.c" />
<ClCompile Include="src\dllmain.cpp" />
</ItemGroup>
<ItemGroup>
<ClInclude Include="src\crc\crc.h" />
<ClInclude Include="src\crc\sniptype.h" />
<ClInclude Include="src\framework.h" />
<ClInclude Include="src\glStuff.h" />
</ItemGroup>
@@ -18,6 +18,9 @@
<ClCompile Include="src\dllmain.cpp">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="src\crc\crc_32.c">
<Filter>Source Files</Filter>
</ClCompile>
</ItemGroup>
<ItemGroup>
<ClInclude Include="src\framework.h">
@@ -26,5 +29,11 @@
<ClInclude Include="src\glStuff.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="src\crc\sniptype.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="src\crc\crc.h">
<Filter>Header Files</Filter>
</ClInclude>
</ItemGroup>
</Project>
@@ -0,0 +1,56 @@
/*
** CRC.H - header file for SNIPPETS CRC and checksum functions
*/
#ifndef CRC__H
#define CRC__H
#include <stdlib.h> /* For size_t */
#include "sniptype.h" /* For BYTE, WORD, DWORD */
/*
** File: ARCCRC16.C
*/
void init_crc_table(void);
WORD crc_calc(WORD crc, char *buf, unsigned nbytes);
void do_file(char *fn);
/*
** File: CRC-16.C
*/
WORD crc16(char *data_p, WORD length);
/*
** File: CRC-16F.C
*/
WORD updcrc(WORD icrc, BYTE *icp, size_t icnt);
/*
** File: CRC_32.C
*/
#define UPDC32(octet,crc) (crc_32_tab[((crc)\
^ ((BYTE)octet)) & 0xff] ^ ((crc) >> 8))
DWORD updateCRC32(unsigned char ch, DWORD crc);
Boolean_T crc32file(char *name, DWORD *crc, long *charcnt);
DWORD crc32buf(char *buf, size_t len);
/*
** File: CHECKSUM.C
*/
unsigned checksum(void *buffer, size_t len, unsigned int seed);
/*
** File: CHECKEXE.C
*/
void checkexe(char *fname);
#endif /* CRC__H */
@@ -0,0 +1,188 @@
/* Crc - 32 BIT ANSI X3.66 CRC checksum files */
#include <stdio.h>
#include "crc.h"
#ifdef __TURBOC__
#pragma warn -cln
#endif
/**********************************************************************\
|* Demonstration program to compute the 32-bit CRC used as the frame *|
|* check sequence in ADCCP (ANSI X3.66, also known as FIPS PUB 71 *|
|* and FED-STD-1003, the U.S. versions of CCITT's X.25 link-level *|
|* protocol). The 32-bit FCS was added via the Federal Register, *|
|* 1 June 1982, p.23798. I presume but don't know for certain that *|
|* this polynomial is or will be included in CCITT V.41, which *|
|* defines the 16-bit CRC (often called CRC-CCITT) polynomial. FIPS *|
|* PUB 78 says that the 32-bit FCS reduces otherwise undetected *|
|* errors by a factor of 10^-5 over 16-bit FCS. *|
\**********************************************************************/
/* Need an unsigned type capable of holding 32 bits; */
typedef DWORD UNS_32_BITS;
/* Copyright (C) 1986 Gary S. Brown. You may use this program, or
code or tables extracted from it, as desired without restriction.*/
/* First, the polynomial itself and its table of feedback terms. The */
/* polynomial is */
/* X^32+X^26+X^23+X^22+X^16+X^12+X^11+X^10+X^8+X^7+X^5+X^4+X^2+X^1+X^0 */
/* Note that we take it "backwards" and put the highest-order term in */
/* the lowest-order bit. The X^32 term is "implied"; the LSB is the */
/* X^31 term, etc. The X^0 term (usually shown as "+1") results in */
/* the MSB being 1. */
/* Note that the usual hardware shift register implementation, which */
/* is what we're using (we're merely optimizing it by doing eight-bit */
/* chunks at a time) shifts bits into the lowest-order term. In our */
/* implementation, that means shifting towards the right. Why do we */
/* do it this way? Because the calculated CRC must be transmitted in */
/* order from highest-order term to lowest-order term. UARTs transmit */
/* characters in order from LSB to MSB. By storing the CRC this way, */
/* we hand it to the UART in the order low-byte to high-byte; the UART */
/* sends each low-bit to hight-bit; and the result is transmission bit */
/* by bit from highest- to lowest-order term without requiring any bit */
/* shuffling on our part. Reception works similarly. */
/* The feedback terms table consists of 256, 32-bit entries. Notes: */
/* */
/* 1. The table can be generated at runtime if desired; code to do so */
/* is shown later. It might not be obvious, but the feedback */
/* terms simply represent the results of eight shift/xor opera- */
/* tions for all combinations of data and CRC register values. */
/* */
/* 2. The CRC accumulation logic is the same for all CRC polynomials, */
/* be they sixteen or thirty-two bits wide. You simply choose the */
/* appropriate table. Alternatively, because the table can be */
/* generated at runtime, you can start by generating the table for */
/* the polynomial in question and use exactly the same "updcrc", */
/* if your application needn't simultaneously handle two CRC */
/* polynomials. (Note, however, that XMODEM is strange.) */
/* */
/* 3. For 16-bit CRCs, the table entries need be only 16 bits wide; */
/* of course, 32-bit entries work OK if the high 16 bits are zero. */
/* */
/* 4. The values must be right-shifted by eight bits by the "updcrc" */
/* logic; the shift must be unsigned (bring in zeroes). On some */
/* hardware you could probably optimize the shift in assembler by */
/* using byte-swap instructions. */
static UNS_32_BITS crc_32_tab[] = { /* CRC polynomial 0xedb88320 */
0x00000000, 0x77073096, 0xee0e612c, 0x990951ba, 0x076dc419, 0x706af48f,
0xe963a535, 0x9e6495a3, 0x0edb8832, 0x79dcb8a4, 0xe0d5e91e, 0x97d2d988,
0x09b64c2b, 0x7eb17cbd, 0xe7b82d07, 0x90bf1d91, 0x1db71064, 0x6ab020f2,
0xf3b97148, 0x84be41de, 0x1adad47d, 0x6ddde4eb, 0xf4d4b551, 0x83d385c7,
0x136c9856, 0x646ba8c0, 0xfd62f97a, 0x8a65c9ec, 0x14015c4f, 0x63066cd9,
0xfa0f3d63, 0x8d080df5, 0x3b6e20c8, 0x4c69105e, 0xd56041e4, 0xa2677172,
0x3c03e4d1, 0x4b04d447, 0xd20d85fd, 0xa50ab56b, 0x35b5a8fa, 0x42b2986c,
0xdbbbc9d6, 0xacbcf940, 0x32d86ce3, 0x45df5c75, 0xdcd60dcf, 0xabd13d59,
0x26d930ac, 0x51de003a, 0xc8d75180, 0xbfd06116, 0x21b4f4b5, 0x56b3c423,
0xcfba9599, 0xb8bda50f, 0x2802b89e, 0x5f058808, 0xc60cd9b2, 0xb10be924,
0x2f6f7c87, 0x58684c11, 0xc1611dab, 0xb6662d3d, 0x76dc4190, 0x01db7106,
0x98d220bc, 0xefd5102a, 0x71b18589, 0x06b6b51f, 0x9fbfe4a5, 0xe8b8d433,
0x7807c9a2, 0x0f00f934, 0x9609a88e, 0xe10e9818, 0x7f6a0dbb, 0x086d3d2d,
0x91646c97, 0xe6635c01, 0x6b6b51f4, 0x1c6c6162, 0x856530d8, 0xf262004e,
0x6c0695ed, 0x1b01a57b, 0x8208f4c1, 0xf50fc457, 0x65b0d9c6, 0x12b7e950,
0x8bbeb8ea, 0xfcb9887c, 0x62dd1ddf, 0x15da2d49, 0x8cd37cf3, 0xfbd44c65,
0x4db26158, 0x3ab551ce, 0xa3bc0074, 0xd4bb30e2, 0x4adfa541, 0x3dd895d7,
0xa4d1c46d, 0xd3d6f4fb, 0x4369e96a, 0x346ed9fc, 0xad678846, 0xda60b8d0,
0x44042d73, 0x33031de5, 0xaa0a4c5f, 0xdd0d7cc9, 0x5005713c, 0x270241aa,
0xbe0b1010, 0xc90c2086, 0x5768b525, 0x206f85b3, 0xb966d409, 0xce61e49f,
0x5edef90e, 0x29d9c998, 0xb0d09822, 0xc7d7a8b4, 0x59b33d17, 0x2eb40d81,
0xb7bd5c3b, 0xc0ba6cad, 0xedb88320, 0x9abfb3b6, 0x03b6e20c, 0x74b1d29a,
0xead54739, 0x9dd277af, 0x04db2615, 0x73dc1683, 0xe3630b12, 0x94643b84,
0x0d6d6a3e, 0x7a6a5aa8, 0xe40ecf0b, 0x9309ff9d, 0x0a00ae27, 0x7d079eb1,
0xf00f9344, 0x8708a3d2, 0x1e01f268, 0x6906c2fe, 0xf762575d, 0x806567cb,
0x196c3671, 0x6e6b06e7, 0xfed41b76, 0x89d32be0, 0x10da7a5a, 0x67dd4acc,
0xf9b9df6f, 0x8ebeeff9, 0x17b7be43, 0x60b08ed5, 0xd6d6a3e8, 0xa1d1937e,
0x38d8c2c4, 0x4fdff252, 0xd1bb67f1, 0xa6bc5767, 0x3fb506dd, 0x48b2364b,
0xd80d2bda, 0xaf0a1b4c, 0x36034af6, 0x41047a60, 0xdf60efc3, 0xa867df55,
0x316e8eef, 0x4669be79, 0xcb61b38c, 0xbc66831a, 0x256fd2a0, 0x5268e236,
0xcc0c7795, 0xbb0b4703, 0x220216b9, 0x5505262f, 0xc5ba3bbe, 0xb2bd0b28,
0x2bb45a92, 0x5cb36a04, 0xc2d7ffa7, 0xb5d0cf31, 0x2cd99e8b, 0x5bdeae1d,
0x9b64c2b0, 0xec63f226, 0x756aa39c, 0x026d930a, 0x9c0906a9, 0xeb0e363f,
0x72076785, 0x05005713, 0x95bf4a82, 0xe2b87a14, 0x7bb12bae, 0x0cb61b38,
0x92d28e9b, 0xe5d5be0d, 0x7cdcefb7, 0x0bdbdf21, 0x86d3d2d4, 0xf1d4e242,
0x68ddb3f8, 0x1fda836e, 0x81be16cd, 0xf6b9265b, 0x6fb077e1, 0x18b74777,
0x88085ae6, 0xff0f6a70, 0x66063bca, 0x11010b5c, 0x8f659eff, 0xf862ae69,
0x616bffd3, 0x166ccf45, 0xa00ae278, 0xd70dd2ee, 0x4e048354, 0x3903b3c2,
0xa7672661, 0xd06016f7, 0x4969474d, 0x3e6e77db, 0xaed16a4a, 0xd9d65adc,
0x40df0b66, 0x37d83bf0, 0xa9bcae53, 0xdebb9ec5, 0x47b2cf7f, 0x30b5ffe9,
0xbdbdf21c, 0xcabac28a, 0x53b39330, 0x24b4a3a6, 0xbad03605, 0xcdd70693,
0x54de5729, 0x23d967bf, 0xb3667a2e, 0xc4614ab8, 0x5d681b02, 0x2a6f2b94,
0xb40bbe37, 0xc30c8ea1, 0x5a05df1b, 0x2d02ef8d
};
DWORD updateCRC32(unsigned char ch, DWORD crc)
{
return UPDC32(ch, crc);
}
Boolean_T crc32file(char *name, DWORD *crc, long *charcnt)
{
FILE *fin;
register DWORD oldcrc32;
register int c;
oldcrc32 = 0xFFFFFFFF; *charcnt = 0;
#ifdef MSDOS
if ((fin=fopen(name, "rb"))==NULL)
#else
if ((fopen_s(&fin, name, "rb"))!=0) // modded for fopen_s to avoid enabling unsafe fopen
#endif
{
perror(name);
return Error_;
}
while ((c=getc(fin))!=EOF)
{
++*charcnt;
oldcrc32 = UPDC32(c, oldcrc32);
}
if (ferror(fin))
{
perror(name);
*charcnt = -1;
}
fclose(fin);
*crc = oldcrc32 = ~oldcrc32;
return Success_;
}
DWORD crc32buf(char *buf, size_t len)
{
register DWORD oldcrc32;
oldcrc32 = 0xFFFFFFFF;
for ( ; len; --len, ++buf)
{
oldcrc32 = UPDC32(*buf, oldcrc32);
}
return ~oldcrc32;
}
#ifdef TEST
main(int argc, char *argv[])
{
DWORD crc;
long charcnt;
register errors = 0;
while(--argc > 0)
{
errors |= crc32file(*++argv, &crc, &charcnt);
printf("%08lX %7ld %s\n", crc, charcnt, *argv);
}
return(errors != 0);
}
#endif /* TEST */
@@ -0,0 +1,81 @@
/*
** SNIPTYPE.H - Include file for SNIPPETS data types and commonly used macros
*/
#ifndef SNIPTYPE__H
#define SNIPTYPE__H
#include <stdlib.h> /* For free() */
#include <string.h> /* For NULL & strlen() */
typedef enum {Error_ = -1, Success_, False_ = 0, True_} Boolean_T;
#if defined(__unix__)
typedef unsigned char BYTE;
typedef unsigned long DWORD;
typedef unsigned short WORD;
#if !defined(FAR)
#define FAR
#endif
#if !defined(NEAR)
#define NEAR
#endif
#if !defined(HUGE)
#define HUGE
#endif
#if !defined(PASCAL)
#define PASCAL
#endif
#if !defined(CDECL)
#define CDECL
#endif
#if !defined(INTERRUPT)
#define INTERRUPT
#endif
#elif !defined(WIN32) && !defined(_WIN32) && !defined(__NT__) \
&& !defined(_WINDOWS)
#if !defined(OS2)
typedef unsigned char BYTE;
typedef unsigned long DWORD;
#endif
typedef unsigned short WORD;
#else
#define WIN32_LEAN_AND_MEAN
#define NOGDI
#define NOSERVICE
#undef INC_OLE1
#undef INC_OLE2
#include <windows.h>
#define HUGE
#endif
typedef union {
signed char c;
BYTE b;
} VAR8_;
typedef union {
VAR8_ v8[2];
signed short s;
WORD w;
} VAR16_;
typedef union {
VAR16_ v16[2];
signed long l;
DWORD dw;
float f;
void *p;
} VAR32_;
typedef union {
VAR32_ v32[2];
double d;
} VAR64_;
#define NUL '\0'
#define LAST_CHAR(s) (((char *)s)[strlen(s) - 1])
#define TOBOOL(x) (!(!(x)))
#define FREE(p) (free(p),(p)=NULL)
#endif /* SNIPTYPE__H */
@@ -2,6 +2,15 @@
#include "glStuff.h"
#include <PluginConfigApi.h>
extern "C" {
// well, only good for 64 bit now, ont that is matters lol
#define SIZEOF_SIZE_T 8
#define SIZEOF_UNSIGNED_LONG_LONG 8
#include <xdelta3.h>
#include <xdelta3.c>
#include "crc/crc.h"
}
// upload to the SSBO instead after processing
// two versions because apparently TexSubImage can cause major stuttering
@@ -185,6 +194,156 @@ void h_glutSetCursor(int cursor)
glutSetCursor(cursor);
}
int64_t hookedGetFileSize(MsString* path) {
// store original size for use later or to return in case of error
int64_t ogsize = divaGetFileSize(path);
std::string pathStr = std::string(path->GetCharBuf());
static std::string shdfarc = "shader.farc";
// identify whether current file is shader or not
bool fileIsShader = false;
if (pathStr.length() >= shdfarc.length())
{
fileIsShader = (0 == pathStr.compare(pathStr.length() - shdfarc.length(), shdfarc.length(), shdfarc));
}
if (!fileIsShader)
{
return ogsize;
}
// reset some stuff when reprocessing shader farc (should never happen)
if (shader_farc_data) {
free(shader_farc_data);
shader_farc_data = NULL;
shader_farc_data_size = 0;
shader_farc_path = "";
}
// if shader, open it and read data
FILE* ogfile;
if (fopen_s(&ogfile, path->GetCharBuf(), "rb") != 0)
{
MessageBoxW(NULL, L"Error opening shader.farc.", L"Novidia", NULL);
return ogsize;
}
void* ogdata = malloc(ogsize);
if (fread(ogdata, 1, ogsize, ogfile) != ogsize)
{
MessageBoxW(NULL, L"Error reading shader.farc.", L"Novidia", NULL);
free(ogdata);
return ogsize;
}
fclose(ogfile);
// crc32 the shader data to find a vcdiff patch file
DWORD crc = crc32buf((char*)ogdata, ogsize);
wchar_t crcstr[9];
swprintf_s(crcstr, L"%x", crc);
//MessageBoxW(NULL, crcstr, L"Novidia", NULL);
std::wstring patch_path = DirPath() + L"\\plugins\\Novidia Shaders\\" + crcstr + L".vcdiff";
DWORD dwAttrib = GetFileAttributesW(patch_path.c_str());
if (dwAttrib == INVALID_FILE_ATTRIBUTES || (dwAttrib & FILE_ATTRIBUTE_DIRECTORY))
{
MessageBoxW(NULL, L"Unable to find shader vcdiff file.", L"Novidia", NULL);
MessageBoxW(NULL, patch_path.c_str(), L"Novidia", NULL);
free(ogdata);
return ogsize; // no patch file exists so can't patch
}
// try opening and reading said vcdiff patch
FILE* vcdfile;
if (_wfopen_s(&vcdfile, patch_path.c_str(), L"rb") != 0)
{
MessageBoxW(NULL, L"Error opening shader vcdiff file.", L"Novidia", NULL);
MessageBoxW(NULL, patch_path.c_str(), L"Novidia", NULL);
free(ogdata);
return ogsize;
}
// get size of the file from having already opened it, skipping whatever the game does
fseek(vcdfile, 0, SEEK_END); // seek to end of file
long vcdsize = ftell(vcdfile); // get current file pointer
fseek(vcdfile, 0, SEEK_SET); // seek back to beginning of file
void* vcddata = malloc(vcdsize);
if (fread(vcddata, 1, vcdsize, vcdfile) != vcdsize)
{
MessageBoxW(NULL, L"Error reading shader vcdiff file.", L"Novidia", NULL);
MessageBoxW(NULL, patch_path.c_str(), L"Novidia", NULL);
free(ogdata);
free(vcddata);
return ogsize;
}
fclose(vcdfile);
// allocate an output buffer and patch shader into it (needs to be done now to know correct size)
usize_t outbuf_size = 64 * 1024 * 1024; // 64M should be enough
void* outbuf = malloc(outbuf_size);
usize_t outbuf_data_size;
int xd3err = xd3_decode_memory((uint8_t*)vcddata, vcdsize, (uint8_t*)ogdata, ogsize, (uint8_t*)outbuf, &outbuf_data_size, outbuf_size, 0);
if (xd3err != 0)
{
MessageBoxW(NULL, L"Error applying shader vcdiff file.", L"Novidia", NULL);
MessageBoxA(NULL, xd3_strerror(xd3err), "Novidia", NULL);
free(ogdata);
free(vcddata);
free(outbuf);
return ogsize;
}
// done successfully! cleanup and save results
free(ogdata);
free(vcddata);
shader_farc_data = outbuf;
shader_farc_data_size = outbuf_data_size;
shader_farc_path = pathStr;
return outbuf_data_size;
}
FILE* hookedFsopen(char* path, char* mode, int* shflag)
{
FILE* res = divaFsopen(path, mode, shflag);
if (strcmp(path, shader_farc_path.c_str()) == 0)
{
shader_file_handle = res;
}
return res;
}
int64_t hookedFread(void* dst, int64_t size, int64_t count, FILE* file)
{
if (file != shader_file_handle)
{
return divaFread(dst, size, count, file);
}
shader_file_handle = NULL; // easy way to avoid chance for reuse
size_t size_bytes = size * count;
if (size_bytes > shader_farc_data_size)
{
return divaFread(dst, size, count, file); // this is an error -- the game wants more data than we have for some reason
}
memcpy(dst, shader_farc_data, size_bytes);
return size_bytes;
}
BOOL APIENTRY DllMain(HMODULE hModule, DWORD ul_reason_for_call, LPVOID lpReserved)
{
@@ -215,6 +374,14 @@ BOOL APIENTRY DllMain(HMODULE hModule, DWORD ul_reason_for_call, LPVOID lpReserv
DetourAttach(&(PVOID&)uploadModelTransformBuf, h_uploadModelTransformBuf_NoUpload);
}
printf("[Novidia] Hooking divaGetFileSize\n");
DetourAttach(&(PVOID&)divaGetFileSize, hookedGetFileSize);
printf("[Novidia] Hooking divaFsopen\n");
DetourAttach(&(PVOID&)divaFsopen, hookedFsopen);
printf("[Novidia] Hooking divaFread\n");
DetourAttach(&(PVOID&)divaFread, hookedFread);
DetourTransactionCommit();
if (shader_amd_farc)
@@ -30,6 +30,46 @@ bool force_BGRA_upload;
GLint tex_upload_format;
bool shader_amd_farc;
struct MsString {
union {
char* string_ptr;
char string_buf[16];
};
uint64_t len;
uint64_t bufsize;
char* GetCharBuf()
{
if (bufsize > 0xf && string_ptr != nullptr)
return string_ptr;
else
return string_buf;
};
void SetCharBuf(char* newcontent)
{
len = strlen(newcontent);
bufsize = len;
if (len > 0xf)
{
string_ptr = _strdup(newcontent);
}
else
{
strcpy_s(string_buf, newcontent);
}
}
};
std::string shader_farc_path = "";
void* shader_farc_data = NULL;
int64_t shader_farc_data_size;
FILE* shader_file_handle = NULL;
int64_t(*divaGetFileSize)(MsString* path) = (int64_t(*)(MsString* path))0x1400abb20;
FILE* (*divaFsopen)(char* path, char* mode, int* shflag) = (FILE* (*)(char* path, char* mode, int* shflag))0x14085a17c;
int64_t(*divaFread)(void* dst, int64_t size, int64_t count, FILE* file) = (int64_t(*)(void* dst, int64_t size, int64_t count, FILE* file))0x14085a6c4;
std::wstring ExePath() {
WCHAR buffer[MAX_PATH];
GetModuleFileNameW(NULL, buffer, MAX_PATH);