1 files changed, 152 insertions, 0 deletions
diff --git a/trafgen_lexer.l b/trafgen_lexer.l
new file mode 100644
index 0000000..69caa58
--- /dev/null
+++ b/trafgen_lexer.l
@@ -0,0 +1,152 @@
+/*
+ * netsniff-ng - the packet sniffing beast
+ * By Daniel Borkmann <daniel@netsniff-ng.org>
+ * Copyright 2012 Daniel Borkmann <dborkma@tik.ee.ethz.ch>,
+ * Swiss federal institute of technology (ETH Zurich)
+ * Subject to the GPL, version 2.
+ */
+
+/* lex-func-prefix: yy */
+
+%{
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <ctype.h>
+
+#include "trafgen_parser.tab.h"
+#include "xmalloc.h"
+
+extern void yyerror(const char *);
+
+static char *try_convert_shellcode(char *sstr)
+{
+	int j = 0;
+	char *bstr, *ostr = sstr, *hay;
+	size_t blen, slen = strlen(sstr), tot = 0;
+	const char *needle = "\\x";
+
+	sstr++;
+	slen -= 2;
+
+	if (slen % 4 != 0)
+		return sstr;
+
+	blen = slen / 4;
+	hay = sstr;
+	while ((hay = strstr(hay, needle)) != NULL ) {
+		hay += strlen(needle) + 2;
+		tot++;
+	}
+
+	if (blen != tot) {
+		printf("Warning: mixed shellcode with strings, "
+		       "using strings!\n");
+		return sstr;
+	}
+
+	blen += 2;
+	bstr = xzmalloc(blen);
+
+	bstr[j++] = '\"';
+	while (j < blen - 1)
+		bstr[j++] = (uint8_t) strtoul(sstr + 2, &sstr, 16);
+	bstr[j++] = '\"';
+
+	xfree(ostr);
+	return bstr;
+}
+
+%}
+
+%option align
+%option nounput
+%option noyywrap
+%option noreject
+%option 8bit
+%option caseless
+%option noinput
+%option nodefault
+
+number_oct	([0][0-9]+)
+number_hex	([0][x][a-fA-F0-9]+)
+number_bin	([0][b][0-1]+)
+number_dec	(([0])|([-+]?[1-9][0-9]*))
+number_ascii	([a-zA-Z])
+
+%%
+
+"cpu"		{ return K_CPU; }
+"fill"		{ return K_FILL; }
+"rnd"		{ return K_RND; }
+"csumip"	{ return K_CSUMIP; }
+"csumip4"	{ return K_CSUMIP; }
+"csumudp"	{ return K_CSUMUDP; }
+"csumtcp"	{ return K_CSUMTCP; }
+"drnd"		{ return K_DRND; }
+"dinc"		{ return K_DINC; }
+"ddec"		{ return K_DDEC; }
+"seqinc"	{ return K_SEQINC; }
+"seqdec"	{ return K_SEQDEC; }
+"const8"|"c8"	{ return K_CONST8; }
+"const16"|"c16"	{ return K_CONST16; }
+"const32"|"c32"	{ return K_CONST32; }
+"const64"|"c64"	{ return K_CONST64; }
+
+[ ]*"-"[ ]*	{ return '-'; }
+[ ]*"+"[ ]*	{ return '+'; }
+[ ]*"*"[ ]*	{ return '*'; }
+[ ]*"/"[ ]*	{ return '/'; }
+[ ]*"%"[ ]*	{ return '%'; }
+[ ]*"&"[ ]*	{ return '&'; }
+[ ]*"|"[ ]*	{ return '|'; }
+[ ]*"<"[ ]*	{ return '<'; }
+[ ]*">"[ ]*	{ return '>'; }
+[ ]*"^"[ ]*	{ return '^'; }
+"{"		{ return '{'; }
+"}"		{ return '}'; }
+"("		{ return '('; }
+")"		{ return ')'; }
+"["		{ return '['; }
+"]"		{ return ']'; }
+","		{ return ','; }
+":"		{ return ':'; }
+
+"\n"		{ yylineno++; }
+
+"\""[^\"]+"\""	{ yylval.str = try_convert_shellcode(xstrdup(yytext));
+		  return string; }
+
+([ \t\r\n]+)?	{ return K_WHITE; }
+
+"/*"([^\*]|\*[^/])*"*/" { return K_COMMENT; }
+
+"#"[^\n]*	{ return K_COMMENT; }
+
+{number_hex}	{ yylval.number = strtoul(yytext, NULL, 16);
+		  return number; }
+
+{number_dec}	{ yylval.number = strtol(yytext, NULL, 10);
+		  return number; }
+
+{number_oct}	{ yylval.number = strtol(yytext + 1, NULL, 8);
+		  return number; }
+
+{number_bin}	{ yylval.number = strtol(yytext + 2, NULL, 2);
+		  return number; }
+
+{number_ascii}	{ yylval.number = (uint8_t) (*yytext);
+		  return number; }
+
+"'\\x"[a-fA-F0-9]{2}"'" { yylval.number = strtol(yytext + 3, NULL, 16);
+		  return number; }
+
+"'"."'"		{ yylval.number = (uint8_t) (*(yytext + 1));
+		  return number; }
+
+";"[^\n]*	{/* NOP */}
+.		{ printf("Unknown character '%s'", yytext);
+		  yyerror("lex Unknown character"); }
+
+%%