acsmx.c

linux下IDS软件,来源于snort社团.

		    if( !px )
		    {
		    printf("*** Out of memory Initializing Aho Corasick in acsmx.c ****");
		    }
	
		    /* Insert at front of MatchList */ 
		    px->next = acsm->acsmStateTable[s].MatchList;
		    acsm->acsmStateTable[s].MatchList = px;
		}
	    }
	}
    }
  
    /* Clean up the queue */ 
    queue_free (queue);
}


/*
*   Build Deterministic Finite Automata from NFA
*/ 
static void
Convert_NFA_To_DFA (ACSM_STRUCT * acsm) 
{
  int r, s;
  int i;
  QUEUE q, *queue = &q;
  
    /* Init a Queue */ 
    queue_init (queue);
  
    /* Add the state 0 transitions 1st */ 
    for (i = 0; i < ALPHABET_SIZE; i++)
    {
      s = acsm->acsmStateTable[0].NextState[i];
      if (s)
	{
	  queue_add (queue, s);
	}
    }
  
    /* Start building the next layer of transitions */ 
    while (queue_count (queue) > 0)
    {
      r = queue_remove (queue);
      
	/* State is a branch state */ 
	for (i = 0; i < ALPHABET_SIZE; i++)
	{
	  if ((s = acsm->acsmStateTable[r].NextState[i]) != ACSM_FAIL_STATE)
	    {
	      queue_add (queue, s);
	    }
	  else
	    {
	      acsm->acsmStateTable[r].NextState[i] =
		acsm->acsmStateTable[acsm->acsmStateTable[r].FailState].
		NextState[i];
	    }
	}
    }
  
    /* Clean up the queue */ 
    queue_free (queue);
}


/*
*
*/ 
ACSM_STRUCT * acsmNew () 
{
  ACSM_STRUCT * p;
  init_xlatcase ();
  p = (ACSM_STRUCT *) AC_MALLOC (sizeof (ACSM_STRUCT));
  MEMASSERT (p, "acsmNew");
  if (p)
    memset (p, 0, sizeof (ACSM_STRUCT));
  return p;
}


/*
*   Add a pattern to the list of patterns for this state machine
*/ 
int
acsmAddPattern (ACSM_STRUCT * p, unsigned char *pat, int n, int nocase,
		int offset, int depth, void * id, int iid) 
{
  ACSM_PATTERN * plist;
  plist = (ACSM_PATTERN *) AC_MALLOC (sizeof (ACSM_PATTERN));
  MEMASSERT (plist, "acsmAddPattern");
  plist->patrn = (unsigned char *) AC_MALLOC (n);
  ConvertCaseEx (plist->patrn, pat, n);
  plist->casepatrn = (unsigned char *) AC_MALLOC (n);
  memcpy (plist->casepatrn, pat, n);
  plist->n = n;
  plist->nocase = nocase;
  plist->offset = offset;
  plist->depth = depth;
  plist->id = id;
  plist->iid = iid;
  plist->next = p->acsmPatterns;
  p->acsmPatterns = plist;
  return 0;
}


/*
*   Compile State Machine
*/ 
int
acsmCompile (ACSM_STRUCT * acsm) 
{
  int i, k;
  ACSM_PATTERN * plist;
  
    /* Count number of states */ 
    acsm->acsmMaxStates = 1;
  for (plist = acsm->acsmPatterns; plist != NULL; plist = plist->next)
    {
      acsm->acsmMaxStates += plist->n;
    }
  acsm->acsmStateTable =
    (ACSM_STATETABLE *) AC_MALLOC (sizeof (ACSM_STATETABLE) *
				   acsm->acsmMaxStates);
  MEMASSERT (acsm->acsmStateTable, "acsmCompile");
  memset (acsm->acsmStateTable, 0,
	    sizeof (ACSM_STATETABLE) * acsm->acsmMaxStates);
  
    /* Initialize state zero as a branch */ 
    acsm->acsmNumStates = 0;
  
    /* Initialize all States NextStates to FAILED */ 
    for (k = 0; k < acsm->acsmMaxStates; k++)
    {
      for (i = 0; i < ALPHABET_SIZE; i++)
	{
	  acsm->acsmStateTable[k].NextState[i] = ACSM_FAIL_STATE;
	}
    }
  
    /* Add each Pattern to the State Table */ 
    for (plist = acsm->acsmPatterns; plist != NULL; plist = plist->next)
    {
      AddPatternStates (acsm, plist);
    }
  
    /* Set all failed state transitions to return to the 0'th state */ 
    for (i = 0; i < ALPHABET_SIZE; i++)
    {
      if (acsm->acsmStateTable[0].NextState[i] == ACSM_FAIL_STATE)
	{
	  acsm->acsmStateTable[0].NextState[i] = 0;
	}
    }
  
    /* Build the NFA  */ 
    Build_NFA (acsm);
  
    /* Convert the NFA to a DFA */ 
    Convert_NFA_To_DFA (acsm);
  
    /*
      printf ("ACSMX-Max Memory: %d bytes, %d states\n", max_memory,
	     acsm->acsmMaxStates);
     */

  
    //Print_DFA( acsm );

    return 0;
}


static unsigned char Tc[64*1024];

/*
*   Search Text or Binary Data for Pattern matches
*/ 
  int
acsmSearch (ACSM_STRUCT * acsm, unsigned char *Tx, int n,
	    int (*Match) (void *  id, int index, void *data), void *data) 
{
  int state;
  ACSM_PATTERN * mlist;
  unsigned char *Tend;
  ACSM_STATETABLE * StateTable = acsm->acsmStateTable;
  int nfound = 0;
  unsigned char *T;
  int index;
  
  /* Case conversion */ 
  ConvertCaseEx (Tc, Tx, n);
  T = Tc;
  Tend = T + n;
 
  for (state = 0; T < Tend; T++)
    {
      state = StateTable[state].NextState[*T];

      if( StateTable[state].MatchList != NULL )
	{
	  for( mlist=StateTable[state].MatchList; mlist!=NULL;
	       mlist=mlist->next )
	    {
	      index = T - mlist->n + 1 - Tc;
	      if( mlist->nocase )
		{
		  nfound++;
		  if (Match (mlist->id, index, data))
		    return nfound;
		}
	      else
		{
		  if( memcmp (mlist->casepatrn, Tx + index, mlist->n) == 0 )
		    {
		      nfound++;
		      if (Match (mlist->id, index, data))
			return nfound;
		    }
		}
	    }
	}
    }
  return nfound;
}


/*
*   Free all memory
*/ 
  void
acsmFree (ACSM_STRUCT * acsm) 
{
  int i;
  ACSM_PATTERN * mlist, *ilist;
  for (i = 0; i < acsm->acsmMaxStates; i++)
    
    {
      if (acsm->acsmStateTable[i].MatchList != NULL)
	
	{
	  mlist = acsm->acsmStateTable[i].MatchList;
	  while (mlist)
	    
	    {
	      ilist = mlist;
	      mlist = mlist->next;
	      AC_FREE (ilist);
	    }
	}
    }
  AC_FREE (acsm->acsmStateTable);
}

/*
 * 
 */ 
/*
static void Print_DFA( ACSM_STRUCT * acsm )
{
    int k;
    int i;
    int next;
    
    for (k = 0; k < acsm->acsmMaxStates; k++)
    {
      for (i = 0; i < ALPHABET_SIZE; i++)
	{
	  next = acsm->acsmStateTable[k].NextState[i];

	  if( next == 0 || next ==  ACSM_FAIL_STATE )
	  {
           if( isprint(i) )
             printf("%3c->%-5d\t",i,next);
           else
             printf("%3d->%-5d\t",i,next);
	  }
	}
      printf("\n");
    }
    
} 
*/
	

int acsmPrintDetailInfo(ACSM_STRUCT * p)
{
    return 0;
}
	
int acsmPrintSummaryInfo(ACSM_STRUCT *acsm )
{
#ifdef XXXXX
    char * fsa[]={
      "TRIE",
      "NFA",
      "DFA",
    };

    ACSM_STRUCT2 * p = &summary.acsm;

    if( !summary.num_states )
	    return;
    
    printf("+--[Pattern Matcher:Aho-Corasick Summary]----------------------\n");
    printf("| Alphabet Size    : %d Chars\n",p->acsmAlphabetSize);
    printf("| Sizeof State     : %d bytes\n",sizeof(acstate_t));
    printf("| Storage Format   : %s \n",sf[ p->acsmFormat ]);
    printf("| Num States       : %d\n",summary.num_states);
    printf("| Num Transitions  : %d\n",summary.num_transitions);
    printf("| State Density    : %.1f%%\n",100.0*(double)summary.num_transitions/(summary.num_states*p->acsmAlphabetSize));
    printf("| Finite Automatum : %s\n", fsa[p->acsmFSA]);
    if( max_memory < 1024*1024 )
    printf("| Memory           : %.2fKbytes\n", (float)max_memory/1024 );
    else
    printf("| Memory           : %.2fMbytes\n", (float)max_memory/(1024*1024) );
    printf("+-------------------------------------------------------------\n");

#endif
    return 0;
}


#ifdef ACSMX_MAIN
  
/*
*  Text Data Buffer
*/ 
unsigned char text[512];

/* 
*    A Match is found
*/ 
  int
MatchFound (unsigned id, int index, void *data) 
{
  fprintf (stdout, "%s\n", (char *) id);
  return 0;
}


/*
*
*/ 
  int
main (int argc, char **argv) 
{
  int i, nocase = 0;
  ACSM_STRUCT * acsm;
  if (argc < 3)
    
    {
      fprintf (stderr,
		"Usage: acsmx pattern word-1 word-2 ... word-n  -nocase\n");
      exit (0);
    }
  acsm = acsmNew ();
  strcpy (text, argv[1]);
  for (i = 1; i < argc; i++)
    if (strcmp (argv[i], "-nocase") == 0)
      nocase = 1;
  for (i = 2; i < argc; i++)
    
    {
      if (argv[i][0] == '-')
	continue;
      acsmAddPattern (acsm, argv[i], strlen (argv[i]), nocase, 0, 0,
			argv[i], i - 2);
    }
  acsmCompile (acsm);
  acsmSearch (acsm, text, strlen (text), MatchFound, (void *) 0);
  acsmFree (acsm);
  printf ("normal pgm end\n");
  return (0);
}
#endif /*  */