Home | History | Annotate | Download | only in interception
      1 //===-- interception_linux.cc -----------------------------------*- C++ -*-===//
      2 //
      3 //                     The LLVM Compiler Infrastructure
      4 //
      5 // This file is distributed under the University of Illinois Open Source
      6 // License. See LICENSE.TXT for details.
      7 //
      8 //===----------------------------------------------------------------------===//
      9 //
     10 // This file is a part of AddressSanitizer, an address sanity checker.
     11 //
     12 // Windows-specific interception methods.
     13 //===----------------------------------------------------------------------===//
     14 
     15 #ifdef _WIN32
     16 
     17 #include "interception.h"
     18 #include <windows.h>
     19 
     20 namespace __interception {
     21 
     22 bool GetRealFunctionAddress(const char *func_name, uptr *func_addr) {
     23   const char *DLLS[] = {
     24     "msvcr80.dll",
     25     "msvcr90.dll",
     26     "kernel32.dll",
     27     NULL
     28   };
     29   *func_addr = 0;
     30   for (size_t i = 0; *func_addr == 0 && DLLS[i]; ++i) {
     31     *func_addr = (uptr)GetProcAddress(GetModuleHandleA(DLLS[i]), func_name);
     32   }
     33   return (*func_addr != 0);
     34 }
     35 
     36 // FIXME: internal_str* and internal_mem* functions should be moved from the
     37 // ASan sources into interception/.
     38 
     39 static void _memset(void *p, int value, size_t sz) {
     40   for (size_t i = 0; i < sz; ++i)
     41     ((char*)p)[i] = (char)value;
     42 }
     43 
     44 static void _memcpy(void *dst, void *src, size_t sz) {
     45   char *dst_c = (char*)dst,
     46        *src_c = (char*)src;
     47   for (size_t i = 0; i < sz; ++i)
     48     dst_c[i] = src_c[i];
     49 }
     50 
     51 static void WriteJumpInstruction(char *jmp_from, char *to) {
     52   // jmp XXYYZZWW = E9 WW ZZ YY XX, where XXYYZZWW is an offset fromt jmp_from
     53   // to the next instruction to the destination.
     54   ptrdiff_t offset = to - jmp_from - 5;
     55   *jmp_from = '\xE9';
     56   *(ptrdiff_t*)(jmp_from + 1) = offset;
     57 }
     58 
     59 static char *GetMemoryForTrampoline(size_t size) {
     60   // Trampolines are allocated from a common pool.
     61   const int POOL_SIZE = 1024;
     62   static char *pool = NULL;
     63   static size_t pool_used = 0;
     64   if (!pool) {
     65     pool = (char *)VirtualAlloc(NULL, POOL_SIZE, MEM_RESERVE | MEM_COMMIT,
     66                                 PAGE_EXECUTE_READWRITE);
     67     // FIXME: Might want to apply PAGE_EXECUTE_READ access after all the
     68     // interceptors are in place.
     69     if (!pool)
     70       return NULL;
     71     _memset(pool, 0xCC /* int 3 */, POOL_SIZE);
     72   }
     73 
     74   if (pool_used + size > POOL_SIZE)
     75     return NULL;
     76 
     77   char *ret = pool + pool_used;
     78   pool_used += size;
     79   return ret;
     80 }
     81 
     82 // Returns 0 on error.
     83 static size_t RoundUpToInstrBoundary(size_t size, char *code) {
     84   size_t cursor = 0;
     85   while (cursor < size) {
     86     switch (code[cursor]) {
     87       case '\x51':  // push ecx
     88       case '\x52':  // push edx
     89       case '\x53':  // push ebx
     90       case '\x54':  // push esp
     91       case '\x55':  // push ebp
     92       case '\x56':  // push esi
     93       case '\x57':  // push edi
     94       case '\x5D':  // pop ebp
     95         cursor++;
     96         continue;
     97       case '\x6A':  // 6A XX = push XX
     98         cursor += 2;
     99         continue;
    100       case '\xE9':  // E9 XX YY ZZ WW = jmp WWZZYYXX
    101         cursor += 5;
    102         continue;
    103     }
    104     switch (*(unsigned short*)(code + cursor)) {  // NOLINT
    105       case 0xFF8B:  // 8B FF = mov edi, edi
    106       case 0xEC8B:  // 8B EC = mov ebp, esp
    107       case 0xC033:  // 33 C0 = xor eax, eax
    108         cursor += 2;
    109         continue;
    110       case 0x458B:  // 8B 45 XX = mov eax, dword ptr [ebp+XXh]
    111       case 0x5D8B:  // 8B 5D XX = mov ebx, dword ptr [ebp+XXh]
    112       case 0xEC83:  // 83 EC XX = sub esp, XX
    113         cursor += 3;
    114         continue;
    115       case 0xC1F7:  // F7 C1 XX YY ZZ WW = test ecx, WWZZYYXX
    116         cursor += 6;
    117         continue;
    118       case 0x3D83:  // 83 3D XX YY ZZ WW TT = cmp TT, WWZZYYXX
    119         cursor += 7;
    120         continue;
    121     }
    122     switch (0x00FFFFFF & *(unsigned int*)(code + cursor)) {
    123       case 0x24448A:  // 8A 44 24 XX = mov eal, dword ptr [esp+XXh]
    124       case 0x24448B:  // 8B 44 24 XX = mov eax, dword ptr [esp+XXh]
    125       case 0x244C8B:  // 8B 4C 24 XX = mov ecx, dword ptr [esp+XXh]
    126       case 0x24548B:  // 8B 54 24 XX = mov edx, dword ptr [esp+XXh]
    127       case 0x24748B:  // 8B 74 24 XX = mov esi, dword ptr [esp+XXh]
    128       case 0x247C8B:  // 8B 7C 24 XX = mov edi, dword ptr [esp+XXh]
    129         cursor += 4;
    130         continue;
    131     }
    132 
    133     // Unknown instruction!
    134     // FIXME: Unknown instruction failures might happen when we add a new
    135     // interceptor or a new compiler version. In either case, they should result
    136     // in visible and readable error messages. However, merely calling abort()
    137     // leads to an infinite recursion in CheckFailed.
    138     // Do we have a good way to abort with an error message here?
    139     __debugbreak();
    140     return 0;
    141   }
    142 
    143   return cursor;
    144 }
    145 
    146 bool OverrideFunction(uptr old_func, uptr new_func, uptr *orig_old_func) {
    147 #ifdef _WIN64
    148 #error OverrideFunction is not yet supported on x64
    149 #endif
    150   // Function overriding works basically like this:
    151   // We write "jmp <new_func>" (5 bytes) at the beginning of the 'old_func'
    152   // to override it.
    153   // We might want to be able to execute the original 'old_func' from the
    154   // wrapper, in this case we need to keep the leading 5+ bytes ('head')
    155   // of the original code somewhere with a "jmp <old_func+head>".
    156   // We call these 'head'+5 bytes of instructions a "trampoline".
    157   char *old_bytes = (char *)old_func;
    158 
    159   // We'll need at least 5 bytes for a 'jmp'.
    160   size_t head = 5;
    161   if (orig_old_func) {
    162     // Find out the number of bytes of the instructions we need to copy
    163     // to the trampoline and store it in 'head'.
    164     head = RoundUpToInstrBoundary(head, old_bytes);
    165     if (!head)
    166       return false;
    167 
    168     // Put the needed instructions into the trampoline bytes.
    169     char *trampoline = GetMemoryForTrampoline(head + 5);
    170     if (!trampoline)
    171       return false;
    172     _memcpy(trampoline, old_bytes, head);
    173     WriteJumpInstruction(trampoline + head, old_bytes + head);
    174     *orig_old_func = (uptr)trampoline;
    175   }
    176 
    177   // Now put the "jmp <new_func>" instruction at the original code location.
    178   // We should preserve the EXECUTE flag as some of our own code might be
    179   // located in the same page (sic!).  FIXME: might consider putting the
    180   // __interception code into a separate section or something?
    181   DWORD old_prot, unused_prot;
    182   if (!VirtualProtect((void *)old_bytes, head, PAGE_EXECUTE_READWRITE,
    183                       &old_prot))
    184     return false;
    185 
    186   WriteJumpInstruction(old_bytes, (char *)new_func);
    187   _memset(old_bytes + 5, 0xCC /* int 3 */, head - 5);
    188 
    189   // Restore the original permissions.
    190   if (!VirtualProtect((void *)old_bytes, head, old_prot, &unused_prot))
    191     return false;  // not clear if this failure bothers us.
    192 
    193   return true;
    194 }
    195 
    196 }  // namespace __interception
    197 
    198 #endif  // _WIN32
    199