unify string replacement

9 files changed, 118 insertions, 181 deletions

diff --git a/extras/ata_id/ata_id.c b/extras/ata_id/ata_id.c
index a4f73b8dfd..d407b0fad9 100644
--- a/extras/ata_id/ata_id.c
+++ b/extras/ata_id/ata_id.c
@@ -44,38 +44,6 @@ static void log_fn(struct udev *udev, int priority,
 	vsyslog(priority, format, args);
 }
 
-static void set_str(char *to, const char *from, size_t count)
-{
-	size_t i, j, len;
-
-	/* strip trailing whitespace */
-	len = strnlen(from, count);
-	while (len && isspace(from[len-1]))
-		len--;
-
-	/* strip leading whitespace */
-	i = 0;
-	while (isspace(from[i]) && (i < len))
-		i++;
-
-	j = 0;
-	while (i < len) {
-		/* substitute multiple whitespace */
-		if (isspace(from[i])) {
-			while (isspace(from[i]))
-				i++;
-			to[j++] = '_';
-		}
-		/* skip chars */
-		if (from[i] == '/') {
-			i++;
-			continue;
-		}
-		to[j++] = from[i++];
-	}
-	to[j] = '\0';
-}
-
 int main(int argc, char *argv[])
 {
 	struct udev *udev;
@@ -146,9 +114,9 @@ int main(int argc, char *argv[])
 		goto close;
 	}
 
-	set_str(model, (char *) id.model, 40);
-	set_str(serial, (char *) id.serial_no, 20);
-	set_str(revision, (char *) id.fw_rev, 8);
+	udev_util_replace_whitespace((char *) id.model, model, 40);
+	udev_util_replace_whitespace((char *) id.serial_no, serial, 20);
+	udev_util_replace_whitespace((char *) id.fw_rev, revision, 8);
 
 	if (export) {
 		if ((id.config >> 8) & 0x80) {
diff --git a/extras/scsi_id/scsi_id.c b/extras/scsi_id/scsi_id.c
index 45e48601f7..69994dd9d3 100644
--- a/extras/scsi_id/scsi_id.c
+++ b/extras/scsi_id/scsi_id.c
@@ -69,39 +69,7 @@ static void log_fn(struct udev *udev, int priority,
 	vsyslog(priority, format, args);
 }
 
-static void set_str(char *to, const char *from, size_t count)
-{
-	size_t i, j, len;
-
-	/* strip trailing whitespace */
-	len = strnlen(from, count);
-	while (len && isspace(from[len-1]))
-		len--;
-
-	/* strip leading whitespace */
-	i = 0;
-	while (isspace(from[i]) && (i < len))
-		i++;
-
-	j = 0;
-	while (i < len) {
-		/* substitute multiple whitespace */
-		if (isspace(from[i])) {
-			while (isspace(from[i]))
-				i++;
-			to[j++] = '_';
-		}
-		/* skip chars */
-		if (from[i] == '/') {
-			i++;
-			continue;
-		}
-		to[j++] = from[i++];
-	}
-	to[j] = '\0';
-}
-
-static void set_type(char *to, const char *from, size_t len)
+static void set_type(const char *from, char *to, size_t len)
 {
 	int type_num;
 	char *eptr;
@@ -533,11 +501,10 @@ static int set_inq_values(struct udev *udev, struct scsi_id_device *dev_scsi, co
 	if (retval)
 		return retval;
 
-	set_str(vendor_str, dev_scsi->vendor, sizeof(vendor_str));
-	set_str(model_str, dev_scsi->model, sizeof(model_str));
-	set_type(type_str, dev_scsi->type, sizeof(type_str));
-	set_str(revision_str, dev_scsi->revision, sizeof(revision_str));
-
+	udev_util_replace_whitespace(dev_scsi->vendor, vendor_str, sizeof(vendor_str));
+	udev_util_replace_whitespace(dev_scsi->model, model_str, sizeof(model_str));
+	set_type(dev_scsi->type, type_str, sizeof(type_str));
+	udev_util_replace_whitespace(dev_scsi->revision, revision_str, sizeof(revision_str));
 	return 0;
 }
 
@@ -604,9 +571,9 @@ static int scsi_id(struct udev *udev, char *maj_min_dev)
 			printf("ID_VENDOR=%s\n", vendor_str);
 			printf("ID_MODEL=%s\n", model_str);
 			printf("ID_REVISION=%s\n", revision_str);
-			set_str(serial_str, dev_scsi.serial, sizeof(serial_str));
+			udev_util_replace_whitespace(dev_scsi.serial, serial_str, sizeof(serial_str));
 			printf("ID_SERIAL=%s\n", serial_str);
-			set_str(serial_str, serial_short, sizeof(serial_str));
+			udev_util_replace_whitespace(serial_short, serial_str, sizeof(serial_str));
 			printf("ID_SERIAL_SHORT=%s\n", serial_str);
 			printf("ID_TYPE=%s\n", type_str);
 		} else {
diff --git a/extras/usb_id/usb_id.c b/extras/usb_id/usb_id.c
index 1499695c67..6dd1d450fb 100644
--- a/extras/usb_id/usb_id.c
+++ b/extras/usb_id/usb_id.c
@@ -46,44 +46,6 @@ static char instance_str[64];
 static int use_usb_info;
 static int use_num_info;
 
-static void set_str(char *to, const char *from, size_t count)
-{
-	size_t i, j, len;
-
-	/* strip trailing whitespace */
-	len = strnlen(from, count);
-	while (len && isspace(from[len-1]))
-		len--;
-
-	/* strip leading whitespace */
-	i = 0;
-	while (isspace(from[i]) && (i < len))
-		i++;
-
-	j = 0;
-	while (i < len) {
-		/* substitute multiple whitespace */
-		if (isspace(from[i])) {
-			while (isspace(from[i]))
-				i++;
-			to[j++] = '_';
-		}
-		/* Replace '/' with '.' */
-		if (from[i] == '/') {
-			to[j++] = '.';
-			i++;
-			continue;
-		}
-		/* skip non-printable chars */
-		if (!isalnum(from[i]) && !ispunct(from[i])) {
-			i++;
-			continue;
-		}
-		to[j++] = from[i++];
-	}
-	to[j] = '\0';
-}
-
 static void set_usb_iftype(char *to, int if_class_num, size_t len)
 {
 	char *type = "generic";
@@ -274,7 +236,7 @@ static int usb_id(struct udev_device *dev)
 			     udev_device_get_sysname(dev_scsi));
 			goto fallback;
 		}
-		set_str(vendor_str, scsi_vendor, sizeof(vendor_str)-1);
+		udev_util_replace_whitespace(scsi_vendor, vendor_str, sizeof(vendor_str)-1);
 
 		scsi_model = udev_device_get_sysattr_value(dev_scsi, "model");
 		if (!scsi_model) {
@@ -282,7 +244,7 @@ static int usb_id(struct udev_device *dev)
 			     udev_device_get_sysname(dev_scsi));
 			goto fallback;
 		}
-		set_str(model_str, scsi_model, sizeof(model_str)-1);
+		udev_util_replace_whitespace(scsi_model, model_str, sizeof(model_str)-1);
 
 		scsi_type = udev_device_get_sysattr_value(dev_scsi, "type");
 		if (!scsi_type) {
@@ -298,7 +260,7 @@ static int usb_id(struct udev_device *dev)
 			     udev_device_get_sysname(dev_scsi));
 			goto fallback;
 		}
-		set_str(revision_str, scsi_rev, sizeof(revision_str)-1);
+		udev_util_replace_whitespace(scsi_rev, revision_str, sizeof(revision_str)-1);
 
 		/*
 		 * some broken devices have the same identifiers
@@ -322,7 +284,7 @@ fallback:
 			info(udev, "No USB vendor information available\n");
 			return 1;
 		}
-		set_str(vendor_str, usb_vendor, sizeof(vendor_str)-1);
+		udev_util_replace_whitespace(usb_vendor, vendor_str, sizeof(vendor_str)-1);
 	}
 
 	if (model_str[0] == '\0') {
@@ -338,7 +300,7 @@ fallback:
 			dbg(udev, "No USB model information available\n");
 			return 1;
 		}
-		set_str(model_str, usb_model, sizeof(model_str)-1);
+		udev_util_replace_whitespace(usb_model, model_str, sizeof(model_str)-1);
 	}
 
 	if (revision_str[0] == '\0') {
@@ -346,7 +308,7 @@ fallback:
 
 		usb_rev = udev_device_get_sysattr_value(dev_usb, "bcdDevice");
 		if (usb_rev)
-			set_str(revision_str, usb_rev, sizeof(revision_str)-1);
+			udev_util_replace_whitespace(usb_rev, revision_str, sizeof(revision_str)-1);
 	}
 
 	if (serial_str[0] == '\0') {
@@ -354,7 +316,7 @@ fallback:
 
 		usb_serial = udev_device_get_sysattr_value(dev_usb, "serial");
 		if (usb_serial)
-			set_str(serial_str, usb_serial, sizeof(serial_str)-1);
+			udev_util_replace_whitespace(usb_serial, serial_str, sizeof(serial_str)-1);
 	}
 	return 0;
 }
diff --git a/extras/volume_id/vol_id.c b/extras/volume_id/vol_id.c
index b4db5adcd8..df47501a6b 100644
--- a/extras/volume_id/vol_id.c
+++ b/extras/volume_id/vol_id.c
@@ -66,38 +66,6 @@ static void vid_log(int priority, const char *file, int line, const char *format
 	return;
 }
 
-static void set_str(char *to, const char *from, size_t count)
-{
-	size_t i, j, len;
-
-	/* strip trailing whitespace */
-	len = strnlen(from, count);
-	while (len && isspace(from[len-1]))
-		len--;
-
-	/* strip leading whitespace */
-	i = 0;
-	while (isspace(from[i]) && (i < len))
-		i++;
-
-	j = 0;
-	while (i < len) {
-		/* substitute multiple whitespace */
-		if (isspace(from[i])) {
-			while (isspace(from[i]))
-				i++;
-			to[j++] = '_';
-		}
-		/* skip chars */
-		if (from[i] == '/') {
-			i++;
-			continue;
-		}
-		to[j++] = from[i++];
-	}
-	to[j] = '\0';
-}
-
 static int all_probers(volume_id_probe_fn_t probe_fn,
 		       struct volume_id *id, uint64_t off, uint64_t size,
 		       void *data)
@@ -288,12 +256,12 @@ int main(int argc, char *argv[])
 		goto exit;
 	}
 
-	set_str(label_safe, label, sizeof(label_safe));
-	util_replace_chars(label_safe, ALLOWED_CHARS_INPUT);
+	udev_util_replace_whitespace(label, label_safe, sizeof(label_safe));
+	udev_util_replace_chars(label_safe, UDEV_ALLOWED_CHARS_INPUT);
 	volume_id_encode_string(label, label_enc, sizeof(label_enc));
 
-	set_str(uuid_safe, uuid, sizeof(uuid_safe));
-	util_replace_chars(uuid_safe, ALLOWED_CHARS_INPUT);
+	udev_util_replace_whitespace(uuid, uuid_safe, sizeof(uuid_safe));
+	udev_util_replace_chars(uuid_safe, UDEV_ALLOWED_CHARS_INPUT);
 	volume_id_encode_string(uuid, uuid_enc, sizeof(uuid_enc));
 
 	volume_id_encode_string(type, type_enc, sizeof(type_enc));
diff --git a/udev/lib/libudev-private.h b/udev/lib/libudev-private.h
index 2f8b8259b5..5e09188f08 100644
--- a/udev/lib/libudev-private.h
+++ b/udev/lib/libudev-private.h
@@ -159,9 +159,10 @@ extern int udev_queue_export_device_finished(struct udev_queue *udev_queue, stru
 extern int udev_queue_export_device_failed(struct udev_queue *udev_queue, struct udev_device *udev_device);
 
 /* libudev-utils */
-#define UTIL_PATH_SIZE		1024
-#define UTIL_LINE_SIZE		2048
-#define UTIL_NAME_SIZE		512
+#define UTIL_PATH_SIZE				1024
+#define UTIL_LINE_SIZE				2048
+#define UTIL_NAME_SIZE				512
+#define UDEV_ALLOWED_CHARS_INPUT		"/ $%?,"
 extern ssize_t util_get_sys_subsystem(struct udev *udev, const char *syspath, char *subsystem, size_t size);
 extern ssize_t util_get_sys_driver(struct udev *udev, const char *syspath, char *driver, size_t size);
 extern int util_resolve_sys_link(struct udev *udev, char *syspath, size_t size);
@@ -171,5 +172,7 @@ extern size_t util_path_decode(char *s);
 extern void util_remove_trailing_chars(char *path, char c);
 extern size_t util_strlcpy(char *dst, const char *src, size_t size);
 extern size_t util_strlcat(char *dst, const char *src, size_t size);
-extern int util_replace_chars(char *str, const char *white);
+extern int udev_util_replace_whitespace(const char *str, char *to, size_t len);
+extern int udev_util_replace_chars(char *str, const char *white);
+extern int udev_util_encode_string(const char *str, char *str_enc, size_t len);
 #endif
diff --git a/udev/lib/libudev-util.c b/udev/lib/libudev-util.c
index 9353e8eda7..f4e78850f9 100644
--- a/udev/lib/libudev-util.c
+++ b/udev/lib/libudev-util.c
@@ -331,8 +331,47 @@ static int utf8_encoded_valid_unichar(const char *str)
 	return len;
 }
 
+int udev_util_replace_whitespace(const char *str, char *to, size_t len)
+{
+	size_t i, j;
+
+	/* strip trailing whitespace */
+	len = strnlen(str, len);
+	while (len && isspace(str[len-1]))
+		len--;
+
+	/* strip leading whitespace */
+	i = 0;
+	while (isspace(str[i]) && (i < len))
+		i++;
+
+	j = 0;
+	while (i < len) {
+		/* substitute multiple whitespace with a single '_' */
+		if (isspace(str[i])) {
+			while (isspace(str[i]))
+				i++;
+			to[j++] = '_';
+		}
+		to[j++] = str[i++];
+	}
+	to[j] = '\0';
+	return 0;
+}
+
+static int is_whitelisted(char c, const char *white)
+{
+	if ((c >= '0' && c <= '9') ||
+	    (c >= 'A' && c <= 'Z') ||
+	    (c >= 'a' && c <= 'z') ||
+	    strchr("#+-.:=@_", c) != NULL ||
+	    (white != NULL && strchr(white, c) != NULL))
+		return 1;
+	return 0;
+}
+
 /* allow chars in whitelist, plain ascii, hex-escaping and valid utf8 */
-int util_replace_chars(char *str, const char *white)
+int udev_util_replace_chars(char *str, const char *white)
 {
 	size_t i = 0;
 	int replaced = 0;
@@ -340,16 +379,7 @@ int util_replace_chars(char *str, const char *white)
 	while (str[i] != '\0') {
 		int len;
 
-		/* accept whitelist */
-		if (white != NULL && strchr(white, str[i]) != NULL) {
-			i++;
-			continue;
-		}
-
-		/* accept plain ascii char */
-		if ((str[i] >= '0' && str[i] <= '9') ||
-		    (str[i] >= 'A' && str[i] <= 'Z') ||
-		    (str[i] >= 'a' && str[i] <= 'z')) {
+		if (is_whitelisted(str[i], white)) {
 			i++;
 			continue;
 		}
@@ -380,6 +410,49 @@ int util_replace_chars(char *str, const char *white)
 		i++;
 		replaced++;
 	}
-
 	return replaced;
 }
+
+/**
+ * util_encode_string:
+ * @str: input string to be encoded
+ * @str_enc: output string to store the encoded input string
+ * @len: maximum size of the output string, which may be
+ *       four times as long as the input string
+ *
+ * Encode all potentially unsafe characters of a string to the
+ * corresponding hex value prefixed by '\x'.
+ *
+ * Returns: 0 if the entire string was copied, non-zero otherwise.
+ **/
+int udev_util_encode_string(const char *str, char *str_enc, size_t len)
+{
+	size_t i, j;
+
+	if (str == NULL || str_enc == NULL || len == 0)
+		return -1;
+
+	str_enc[0] = '\0';
+	for (i = 0, j = 0; str[i] != '\0'; i++) {
+		int seqlen;
+
+		seqlen = utf8_encoded_valid_unichar(&str[i]);
+		if (seqlen > 1) {
+			memcpy(&str_enc[j], &str[i], seqlen);
+			j += seqlen;
+			i += (seqlen-1);
+		} else if (str[i] == '\\' || !is_whitelisted(str[i], NULL)) {
+			sprintf(&str_enc[j], "\\x%02x", (unsigned char) str[i]);
+			j += 4;
+		} else {
+			str_enc[j] = str[i];
+			j++;
+		}
+		if (j+3 >= len)
+			goto err;
+	}
+	str_enc[j] = '\0';
+	return 0;
+err:
+	return -1;
+}
diff --git a/udev/udev-event.c b/udev/udev-event.c
index fe579b49c7..2d513d7b65 100644
--- a/udev/udev-event.c
+++ b/udev/udev-event.c
@@ -319,7 +319,7 @@ found:
 				size = strlen(value);
 				while (size > 0 && isspace(value[--size]))
 					value[size] = '\0';
-				count = util_replace_chars(value, ALLOWED_CHARS_INPUT);
+				count = udev_util_replace_chars(value, UDEV_ALLOWED_CHARS_INPUT);
 				if (count > 0)
 					info(event->udev, "%i character(s) replaced\n" , count);
 				util_strlcat(string, value, maxsize);
diff --git a/udev/udev-rules.c b/udev/udev-rules.c
index e07cd3264c..876aa76713 100644
--- a/udev/udev-rules.c
+++ b/udev/udev-rules.c
@@ -2085,7 +2085,7 @@ int udev_rules_apply_to_event(struct udev_rules *rules, struct udev_event *event
 
 					util_remove_trailing_chars(result, '\n');
 					if (esc == ESCAPE_UNSET || esc == ESCAPE_REPLACE) {
-						count = util_replace_chars(result, ALLOWED_CHARS_INPUT);
+						count = udev_util_replace_chars(result, UDEV_ALLOWED_CHARS_INPUT);
 						if (count > 0)
 							info(event->udev, "%i character(s) replaced\n" , count);
 					}
@@ -2284,7 +2284,7 @@ int udev_rules_apply_to_event(struct udev_rules *rules, struct udev_event *event
 				util_strlcpy(name_str, name, sizeof(name_str));
 				udev_event_apply_format(event, name_str, sizeof(name_str));
 				if (esc == ESCAPE_UNSET || esc == ESCAPE_REPLACE) {
-					count = util_replace_chars(name_str, ALLOWED_CHARS_FILE);
+					count = udev_util_replace_chars(name_str, "/");
 					if (count > 0)
 						info(event->udev, "%i character(s) replaced\n", count);
 					free(event->name);
@@ -2316,9 +2316,9 @@ int udev_rules_apply_to_event(struct udev_rules *rules, struct udev_event *event
 				util_strlcpy(temp, &rules->buf[cur->key.value_off], sizeof(temp));
 				udev_event_apply_format(event, temp, sizeof(temp));
 				if (esc == ESCAPE_UNSET)
-					count = util_replace_chars(temp, ALLOWED_CHARS_FILE " ");
+					count = udev_util_replace_chars(temp, "/ ");
 				else if (esc == ESCAPE_REPLACE)
-					count = util_replace_chars(temp, ALLOWED_CHARS_FILE);
+					count = udev_util_replace_chars(temp, "/");
 				if (count > 0)
 					info(event->udev, "%i character(s) replaced\n" , count);
 				dbg(event->udev, "rule applied, added symlink(s) '%s'\n", temp);
diff --git a/udev/udev.h b/udev/udev.h
index 732b131390..ea8c8d919d 100644
--- a/udev/udev.h
+++ b/udev/udev.h
@@ -26,10 +26,6 @@
 #include "lib/libudev.h"
 #include "lib/libudev-private.h"
 
-#define ALLOWED_CHARS				"#+-.:=@_"
-#define ALLOWED_CHARS_FILE			ALLOWED_CHARS "/"
-#define ALLOWED_CHARS_INPUT			ALLOWED_CHARS_FILE " $%?,"
-
 #define DEFAULT_FAKE_PARTITIONS_COUNT		15
 #define UDEV_EVENT_TIMEOUT			180