New upstream version 15.3upstream/15.3

1 files changed, 134 insertions, 47 deletions

diff --git a/include/str.h b/include/str.h
index 9a748366..d433e6ec 100644
--- a/include/str.h
+++ b/include/str.h
@@ -1,65 +1,152 @@
+// SPDX-License-Identifier: BSD-2-Clause-Patent
+
 #ifndef SHIM_STR_H
 #define SHIM_STR_H
 
-static inline
-__attribute__((unused))
-unsigned long strnlena(const CHAR8 *s, unsigned long n)
-{
-	unsigned long i;
-	for (i = 0; i <= n; i++)
-		if (s[i] == '\0')
-			break;
-	return i;
-}
+#if __GNUC__ > 6
+#ifdef SHIM_UNIT_TEST
+#pragma GCC diagnostic error "-Wnonnull-compare"
+#else
+#pragma GCC diagnostic ignored "-Wnonnull-compare"
+#endif
+#endif
 
-static inline
-__attribute__((unused))
-CHAR8 *
-strncpya(CHAR8 *dest, const CHAR8 *src, unsigned long n)
+static inline UNUSED RETURNS_NONNULL NONNULL(1)
+char *
+strnchrnul(const char *s, size_t max, int c)
 {
-	unsigned long i;
+	unsigned int i;
+
+	if (!s || !max)
+		return (char *)s;
+
+	for (i = 0; i < max && s[i] != '\0' && s[i] != c; i++)
+		;
 
-	for (i = 0; i < n && src[i] != '\0'; i++)
-		dest[i] = src[i];
-	for (; i < n; i++)
-		dest[i] = '\0';
+	if (i == max)
+		i--;
 
-	return dest;
+	return (char *)&s[i];
 }
 
-static inline
-__attribute__((unused))
-CHAR8 *
-strcata(CHAR8 *dest, const CHAR8 *src)
+/**
+ * strntoken: tokenize a string, with a limit
+ * str: your string (will be modified)
+ * max: maximum number of bytes to ever touch
+ * delims: string of one character delimeters, any of which will tokenize
+ * *token: the token we're passing back (must be a pointer to NULL initially)
+ * state: a pointer to one char of state for between calls
+ *
+ * Ensure that both token and state are preserved across calls.  Do:
+ *   char state = 0;
+ *   char *token = NULL;
+ *   for (...) {
+ *     valid = strntoken(...)
+ * not:
+ *   char state = 0;
+ *   for (...) {
+ *     char *token = NULL;
+ *     valid = strntoken(...)
+ *
+ * - it will not test bytes beyond str[max-1]
+ * - it will not set *token to an address beyond &str[max-1]
+ * - it will set *token to &str[max-1] without testing &str[max-2] for
+ *   &str[max-1] == str
+ * - sequences of multiple delimeters will result in empty (pointer to '\0')
+ *   tokens.
+ * - it expects you to update str and max on successive calls.
+ *
+ * return:
+ * true means it hasn't tested str[max-1] yet and token is valid
+ * false means it got to a NUL or str[max-1] and token is invalid
+ */
+static inline UNUSED NONNULL(1, 3, 4) int
+strntoken(char *str, size_t max, const char *delims, char **token, char *state)
 {
-	unsigned long dest_len = strlena(dest);
-	unsigned long i;
+	char *tokend;
+	const char *delim;
+	int isdelim = 0;
+	int state_is_delim = 0;
 
-	for (i = 0; src[i] != '\0'; i++)
-		dest[dest_len + i] = src[i];
-	dest[dest_len + i] = '\0';
+	if (!str || !max || !delims || !token || !state)
+		return 0;
 
-	return dest;
+	tokend = &str[max-1];
+	if (!str || max == 0 || !delims || !token)
+		return 0;
+
+	/*
+	 * the very special case of "" with max=1, where we have no prior
+	 * state to let us know this is the same as right after a delim
+	 */
+	if (*token == NULL && max == 1 && *str == '\0') {
+		state_is_delim = 1;
+	}
+
+	for (delim = delims; *delim; delim++) {
+		char *tmp = NULL;
+		if (*token && *delim == *state)
+			state_is_delim = 1;
+		tmp = strnchrnul(str, max, *delim);
+		if (tmp < tokend)
+			tokend = tmp;
+		if (*tokend == *delim)
+			isdelim = 1;
+	}
+	*token = str;
+	if (isdelim) {
+		*state = *tokend;
+		*tokend = '\0';
+		return 1;
+	}
+	return state_is_delim;
 }
 
-static inline
-__attribute__((unused))
-CHAR8 *
-translate_slashes(char *str)
+#define UTF8_BOM { 0xef, 0xbb, 0xbf }
+#define UTF8_BOM_SIZE 3
+
+static inline UNUSED NONNULL(1) BOOLEAN
+is_utf8_bom(CHAR8 *buf, size_t bufsize)
 {
-	int i;
-	int j;
-	if (str == NULL)
-		return (CHAR8 *)str;
-
-	for (i = 0, j = 0; str[i] != '\0'; i++, j++) {
-		if (str[i] == '\\') {
-			str[j] = '/';
-			if (str[i+1] == '\\')
-				i++;
-		}
-	}
-	return (CHAR8 *)str;
+	unsigned char bom[] = UTF8_BOM;
+
+	return CompareMem(buf, bom, MIN(UTF8_BOM_SIZE, bufsize)) == 0;
 }
 
+/**
+ * parse CSV data from data to end.
+ * *data	points to the first byte of the data
+ * end		points to a NUL byte at the end of the data
+ * n_columns	number of columns per entry
+ * list		the list head we're adding to
+ *
+ * On success, list will be populated with individually allocate a list of
+ * struct csv_list objects, with one column per entry of the "columns" array,
+ * filled left to right with up to n_columns elements, or NULL when a csv line
+ * does not have enough elements.
+ *
+ * Note that the data will be modified; all comma, linefeed, and newline
+ * characters will be set to '\000'.  Additionally, consecutive linefeed and
+ * newline characters will not result in rows in the results.
+ *
+ * On failure, list will be empty and all entries on it will have been freed,
+ * using free_csv_list(), whether they were there before calling
+ * parse_csv_data or not.
+ */
+
+struct csv_row {
+	list_t list;		/* this is a linked list */
+	size_t n_columns;	/* this is how many columns are actually populated */
+	char *columns[0];	/* these are pointers to columns */
+};
+
+EFI_STATUS parse_csv_data(char *data, char *end, size_t n_columns,
+                          list_t *list);
+void free_csv_list(list_t *list);
+
+#ifdef SHIM_UNIT_TEST
+void NONNULL(1, 3, 4)
+parse_csv_line(char * line, size_t max, size_t *n_columns, const char *columns[]);
+#endif
+
 #endif /* SHIM_STR_H */