compiler: parse port numbers in ranges using the standard number parser.

Run them through the same code that's used for numbers in the lexical
analyzer, for consistency.

1 files changed, 131 insertions, 38 deletions

diff --git a/scanner.l b/scanner.l
index 85fe395a..c20637b2 100644
--- a/scanner.l
+++ b/scanner.l
@@ -32,6 +32,26 @@
 #include "grammar.h"
 
 #include "diag-control.h"
+
+/*
+ * Convert string to 32-bit unsigned integer; the string starts at
+ * string and is string_len bytes long.
+ *
+ * On success, sets *val to the value and returns 1.
+ * On failure, sets the BPF error string and returns 0.
+ *
+ * Also used in gencode.c
+ */
+typedef enum {
+	STOULEN_OK,
+	STOULEN_NOT_HEX_NUMBER,
+	STOULEN_NOT_OCTAL_NUMBER,
+	STOULEN_NOT_DECIMAL_NUMBER,
+	STOULEN_ERROR
+} stoulen_ret;
+
+stoulen_ret stoulen(const char *string, size_t stringlen, bpf_u_int32 *val,
+    compiler_state_t *cstate);
 }
 
 /*
@@ -149,7 +169,7 @@ void pcap_set_column(int, yyscan_t);
 #include "os-proto.h"
 #endif
 
-static int stou(char *, YYSTYPE *, compiler_state_t *);
+static int stou(const char *, YYSTYPE *, compiler_state_t *);
 
 /*
  * Disable diagnostics in the code generated by Flex.
@@ -490,27 +510,20 @@ tcp-cwr			{ yylval->h = 0x80; return NUM; }
  */
 DIAG_ON_FLEX
 
-/*
- * Convert string to 32-bit unsigned integer.  Just like atoi(), but checks for
- * preceding 0x or 0 and uses hex or octal instead of decimal.
- *
- * On success, sets yylval->h to the value and returns NUM.
- * On failure, sets the BPF error string and returns LEX_ERROR, to force
- * the parse to stop.
- */
-static int
-stou(char *yytext_arg, YYSTYPE *yylval_arg, compiler_state_t *yyextra_arg)
+stoulen_ret
+stoulen(const char *string, size_t string_len, bpf_u_int32 *val,
+    compiler_state_t *cstate)
 {
 	bpf_u_int32 n = 0;
 	unsigned int digit;
-	char *s = yytext_arg;
+	const char *s = string;
 
 	/*
-	 * yytext_arg is guaranteed either to be a string of decimal digits
+	 * string is guaranteed either to be a string of decimal digits
 	 * or 0[xX] followed by a string of hex digits.
 	 */
-	if (*s == '0') {
-		if (s[1] == 'x' || s[1] == 'X') {
+	if (string_len >= 1 && *s == '0') {
+		if (string_len >= 2  && (s[1] == 'x' || s[1] == 'X')) {
 			/*
 			 * Begins with 0x or 0X, so hex.
 			 * Guaranteed to be all hex digits following the
@@ -518,13 +531,25 @@ stou(char *yytext_arg, YYSTYPE *yylval_arg, compiler_state_t *yyextra_arg)
 			 * A-F.
 			 */
 			s += 2;	/* skip the prefix */
-			while ((digit = *s++) != '\0') {
+			string_len -= 2;
+			while (string_len != 0) {
+				digit = *s++;
+				string_len--;
 				if (digit >= '0' && digit <= '9')
 					digit = digit - '0';
 				else if (digit >= 'a' && digit <= 'f')
 					digit = digit - 'a' + 10;
-				else
+				else if (digit >= 'A' && digit <= 'F')
 					digit = digit - 'A' + 10;
+				else {
+					/*
+					 * Not a valid hex number.
+					 * Don't treat this as an error,
+					 * in case the caller wants to
+					 * interpret it as something else.
+					 */
+					return STOULEN_NOT_HEX_NUMBER;
+				}
 
 				/*
 				 * Check for overflow.
@@ -536,10 +561,10 @@ stou(char *yytext_arg, YYSTYPE *yylval_arg, compiler_state_t *yyextra_arg)
 					 * add 4 more; that won't fit
 					 * in 32 bits.
 					 */
-					bpf_set_error(yyextra_arg,
-					    "number %s overflows 32 bits",
-					    yytext_arg);
-					return LEX_ERROR;
+					bpf_set_error(cstate,
+					    "number %.*s overflows 32 bits",
+					   (int)string_len, string);
+					return STOULEN_ERROR;
 				}
 				n = (n << 4) + digit;
 			}
@@ -551,14 +576,20 @@ stou(char *yytext_arg, YYSTYPE *yylval_arg, compiler_state_t *yyextra_arg)
 			 * report an error.
 			 */
 			s += 1;
-			while ((digit = *s++) != '\0') {
+			string_len -= 1;
+			while (string_len != 0) {
+				digit = *s++;
+				string_len--;
 				if (digit >= '0' && digit <= '7')
 					digit = digit - '0';
 				else {
-					bpf_set_error(yyextra_arg,
-					    "number %s contains non-octal digit",
-					    yytext_arg);
-					return LEX_ERROR;
+					/*
+					 * Not a valid octal number.
+					 * Don't treat this as an error,
+					 * in case the caller wants to
+					 * interpret it as something else.
+					 */
+					return STOULEN_NOT_OCTAL_NUMBER;
 				}
 				if (n > 03777777777U) {
 					/*
@@ -567,10 +598,10 @@ stou(char *yytext_arg, YYSTYPE *yylval_arg, compiler_state_t *yyextra_arg)
 					 * 3 more; that won't fit in
 					 * 32 bits.
 					 */
-					bpf_set_error(yyextra_arg,
-					    "number %s overflows 32 bits",
-					    yytext_arg);
-					return LEX_ERROR;
+					bpf_set_error(cstate,
+					    "number %.*s overflows 32 bits",
+					   (int)string_len, string);
+					return STOULEN_ERROR;
 				}
 				n = (n << 3) + digit;
 			}
@@ -579,21 +610,83 @@ stou(char *yytext_arg, YYSTYPE *yylval_arg, compiler_state_t *yyextra_arg)
 		/*
 		 * Decimal.
 		 */
-		while ((digit = *s++) != '\0') {
-			digit = digit - '0';
+		while (string_len != 0) {
+			digit = *s++;
+			string_len--;
+			if (digit >= '0' && digit <= '9')
+				digit = digit - '0';
+			else {
+				/*
+				 * Not a valid decimal number.
+				 * Don't treat this as an error,
+				 * in case the caller wants to
+				 * interpret it as something else.
+				 */
+				return STOULEN_NOT_DECIMAL_NUMBER;
+			}
 #define CUTOFF_DEC	(0xFFFFFFFFU / 10U)
 #define CUTLIM_DEC	(0xFFFFFFFFU % 10U)
 			if (n > CUTOFF_DEC ||
 			    (n == CUTOFF_DEC && digit > CUTLIM_DEC)) {
-				bpf_set_error(yyextra_arg,
-				    "number %s overflows 32 bits",
-				    yytext_arg);
-				return LEX_ERROR;
+				/*
+				 * Adding that digit will result in a
+				 * number that won't fit in 32 bits.
+				 */
+				bpf_set_error(cstate,
+				    "number %.*s overflows 32 bits",
+				   (int)string_len, string);
+				return STOULEN_ERROR;
 			}
 			n = (n * 10) + digit;
 		}
 	}
 
-	yylval_arg->h = n;
-	return NUM;
+	*val = n;
+	return STOULEN_OK;
+}
+
+/*
+ * Convert string to 32-bit unsigned integer.  Just like atoi(), but checks for
+ * preceding 0x or 0 and uses hex or octal instead of decimal.
+ *
+ * On success, sets yylval->h to the value and returns NUM.
+ * On failure, sets the BPF error string and returns LEX_ERROR, to force
+ * the parse to stop.
+ */
+static int
+stou(const char *yytext_arg, YYSTYPE *yylval_arg, compiler_state_t *yyextra_arg)
+{
+	stoulen_ret ret;
+
+	ret = stoulen(yytext_arg, strlen(yytext_arg), &yylval_arg->h,
+	    yyextra_arg);
+	switch (ret) {
+
+	case STOULEN_OK:
+		return NUM;
+
+	case STOULEN_NOT_OCTAL_NUMBER:
+		bpf_set_error(yyextra_arg, "number %s contains non-octal digit",
+		    yytext_arg);
+		return LEX_ERROR;
+
+	case STOULEN_NOT_HEX_NUMBER:
+		bpf_set_error(yyextra_arg, "number %s contains non-hex digit",
+		    yytext_arg);
+		return LEX_ERROR;
+
+	case STOULEN_NOT_DECIMAL_NUMBER:
+		bpf_set_error(yyextra_arg, "number %s contains non-decimal digit",
+		    yytext_arg);
+		return LEX_ERROR;
+
+	case STOULEN_ERROR:
+		/* Error already set. */
+		return LEX_ERROR;
+
+	default:
+		/* Should not happen */
+		bpf_set_error(yyextra_arg, "stoulen returned %d", ret);
+		return LEX_ERROR;
+	}
 }