kernel: Support quoted values in cmdline arguments with new parser

apps/kcmdline.c

@@ -23,62 +23,7 @@
 #include <unistd.h>
 #include <toaru/hashmap.h>
 
-hashmap_t * args_map = NULL;
+#include "../kernel/misc/args.c"
-
-int tokenize(char * str, char * sep, char **buf) {
-	char * pch_i;
-	char * save_i;
-	int    argc = 0;
-	pch_i = strtok_r(str,sep,&save_i);
-	if (!pch_i) { return 0; }
-	while (pch_i != NULL) {
-		buf[argc] = (char *)pch_i;
-		++argc;
-		pch_i = strtok_r(NULL,sep,&save_i);
-	}
-	buf[argc] = NULL;
-	return argc;
-}
-
-void args_parse(char * _arg) {
-	char * arg = strdup(_arg);
-	char * argv[1024];
-	int argc = tokenize(arg, " ", argv);
-
-	for (int i = 0; i < argc; ++i) {
-		char * c = strdup(argv[i]);
-
-		char * name;
-		char * value;
-
-		name = c;
-		value = NULL;
-		/* Find the first = and replace it with a null */
-		char * v = c;
-		while (*v) {
-			if (*v == '=') {
-				*v = '\0';
-				v++;
-				value = v;
-				char * tmp = value;
-				/* scan it for \037 and replace with spaces */
-				while (*tmp) {
-					if (*tmp == '\037') {
-						*tmp = ' ';
-					}
-					tmp++;
-				}
-				goto _break;
-			}
-			v++;
-		}
-
-_break:
-		hashmap_set(args_map, name, value);
-	}
-
-	free(arg);
-}
 
 void show_usage(int argc, char * argv[]) {
 	printf(
@@ -94,28 +39,15 @@ void show_usage(int argc, char * argv[]) {
 }
 
 int main(int argc, char * argv[]) {
-	/* Open */
+	char * cmdline = args_from_procfs();
-	FILE * f = fopen("/proc/cmdline", "r");
-	if (!f) return 1;
-
-	/* Read */
-	char * cmdline = malloc(4096); /* cmdline can't be longer than that */
-	memset(cmdline, 0, 4096);
-	size_t size = fread(cmdline, 1, 4096, f);
-	if (cmdline[size-1] == '\n') cmdline[size-1] = '\0';
-	fclose(f);
-
-	/* Parse */
-	args_map = hashmap_create(10);
-	args_parse(cmdline);
 
 	/* Figure out what we're doing */
 	int opt;
 	while ((opt = getopt(argc, argv, "?g:q:s")) != -1) {
 		switch (opt) {
 			case 'g':
-				if (hashmap_has(args_map, optarg)) {
+				if (hashmap_has(kernel_args_map, optarg)) {
-					char * tmp = (char*)hashmap_get(args_map, optarg);
+					char * tmp = (char*)hashmap_get(kernel_args_map, optarg);
 					if (!tmp) {
 						printf("%s\n", optarg); /* special case = present but not set should yield name of variable */
 					} else {
@@ -126,9 +58,9 @@ int main(int argc, char * argv[]) {
 					return 1;
 				}
 			case 'q':
-				return !hashmap_has(args_map,optarg);
+				return !hashmap_has(kernel_args_map,optarg);
 			case 's':
-				return size;
+				return strlen(cmdline);
 			case '?':
 				show_usage(argc, argv);
 				return 1;

apps/migrate.c

@@ -37,21 +37,6 @@
 static int _debug = 0;
 static FILE * _splash = NULL;
 
-int tokenize(char * str, char * sep, char **buf) {
-	char * pch_i;
-	char * save_i;
-	int    argc = 0;
-	pch_i = strtok_r(str,sep,&save_i);
-	if (!pch_i) { return 0; }
-	while (pch_i != NULL) {
-		buf[argc] = (char *)pch_i;
-		++argc;
-		pch_i = strtok_r(NULL,sep,&save_i);
-	}
-	buf[argc] = NULL;
-	return argc;
-}
-
 void copy_link(char * source, char * dest, int mode, int uid, int gid) {
 	//fprintf(stderr, "need to copy link %s to %s\n", source, dest);
 	char tmp[1024];
@@ -144,51 +129,11 @@ void free_ramdisk(char * path) {
 	close(fd);
 }
 
-hashmap_t * get_cmdline(void) {
+#include "../kernel/misc/args.c"
-	int fd = open("/proc/cmdline", O_RDONLY);
+static hashmap_t * get_cmdline(void) {
-	char * out = malloc(1024);
+	char * results = args_from_procfs();
-	size_t r = read(fd, out, 1024);
+	if (results) free(results);
-	out[r] = '\0';
+	return kernel_args_map;
-	if (out[r-1] == '\n') {
-		out[r-1] = '\0';
-	}
-
-	char * arg = strdup(out);
-	char * argv[1024];
-	int argc = tokenize(arg, " ", argv);
-
-	/* New let's parse the tokens into the arguments list so we can index by key */
-
-	hashmap_t * args = hashmap_create(10);
-
-	for (int i = 0; i < argc; ++i) {
-		char * c = strdup(argv[i]);
-
-		char * name;
-		char * value;
-
-		name = c;
-		value = NULL;
-		/* Find the first = and replace it with a null */
-		char * v = c;
-		while (*v) {
-			if (*v == '=') {
-				*v = '\0';
-				v++;
-				value = v;
-				goto _break;
-			}
-			v++;
-		}
-
-_break:
-		hashmap_set(args, name, value);
-	}
-
-	free(arg);
-	free(out);
-
-	return args;
 }
 
 int main(int argc, char * argv[]) {

apps/splash-log.c

@@ -67,69 +67,13 @@ static void say_hello(void) {
 	update_message(hello_msg);
 }
 
-static int tokenize(char * str, char * sep, char **buf) {
+#include "../kernel/misc/args.c"
-	char * pch_i;
-	char * save_i;
-	int    argc = 0;
-	pch_i = strtok_r(str,sep,&save_i);
-	if (!pch_i) { return 0; }
-	while (pch_i != NULL) {
-		buf[argc] = (char *)pch_i;
-		++argc;
-		pch_i = strtok_r(NULL,sep,&save_i);
-	}
-	buf[argc] = NULL;
-	return argc;
-}
-
 static hashmap_t * get_cmdline(void) {
-	int fd = open("/proc/cmdline", O_RDONLY);
+	char * results = args_from_procfs();
-	char * out = malloc(1024);
+	if (results) free(results);
-	size_t r = read(fd, out, 1024);
+	return kernel_args_map;
-	out[r] = '\0';
-	if (out[r-1] == '\n') {
-		out[r-1] = '\0';
-	}
-
-	char * arg = strdup(out);
-	char * argv[1024];
-	int argc = tokenize(arg, " ", argv);
-
-	/* New let's parse the tokens into the arguments list so we can index by key */
-
-	hashmap_t * args = hashmap_create(10);
-
-	for (int i = 0; i < argc; ++i) {
-		char * c = strdup(argv[i]);
-
-		char * name;
-		char * value;
-
-		name = c;
-		value = NULL;
-		/* Find the first = and replace it with a null */
-		char * v = c;
-		while (*v) {
-			if (*v == '=') {
-				*v = '\0';
-				v++;
-				value = v;
-				goto _break;
-			}
-			v++;
-		}
-
-_break:
-		hashmap_set(args, name, value);
-	}
-
-	free(arg);
-	free(out);
-
-	return args;
 }
 
-
 int main(int argc, char * argv[]) {
 	if (getuid() != 0) {
 		fprintf(stderr, "%s: only root should run this\n", argv[0]);

boot/config.c

@@ -39,7 +39,7 @@ char * LINK_TEXT = "https://toaruos.org - https://github.com/klange/toaruos";
 /* Boot command line strings */
 #define DEFAULT_ROOT_CMDLINE "root=/dev/ram0 "
 #define DEFAULT_GRAPHICAL_CMDLINE "start=live-session "
-#define DEFAULT_SINGLE_CMDLINE "start=terminal\037-F "
+#define DEFAULT_SINGLE_CMDLINE "start=\"terminal -F\" "
 #define DEFAULT_TEXT_CMDLINE "start=--vga vid=text "
 #define DEFAULT_VID_CMDLINE "vid=auto "
 #define MIGRATE_CMDLINE "migrate "

kernel/misc/args.c

@@ -6,14 +6,34 @@
  * hard disk partition should be mounted as root. We parse them
  * into a hash table for easy lookup by key.
  *
+ * An argument may be value-less (having no '='), in which case
+ * its value in the hash table will be NULL but it will be present.
+ * Examples of value-less arguments are @c lfbwc or @c noi965.
+ *
+ * Arguments with values can have quoted or unquoted values. Unquoted
+ * values are terminated by a space or the end of the command line and
+ * are not processed for escapes. Examples of arguments with
+ * unquoted values are @c root=/dev/ram0 or @c start=live-session.
+ *
+ * Quoted values must started immediately with a double quote (").
+ * Double quotes within the value may be escaped with a backslash (\).
+ * Backslash can also be escaped. Any other character after a
+ * backslash results in both a literal backslash and the following
+ * character.
+ *
+ * If a quoted value is not properly terminated with an unescaped
+ * double quote character, the entire argument will be ignored.
+ *
  * @copyright This file is part of ToaruOS and is released under the terms
  *            of the NCSA / University of Illinois License - see LICENSE.md
- *            Copyright (C) 2011-2021 K. Lange
+ *            Copyright (C) 2011-2023 K. Lange
  */
-#include <kernel/string.h>
+#ifdef _KERNEL_
-#include <kernel/args.h>
+# include <kernel/string.h>
-#include <kernel/tokenize.h>
+# include <kernel/args.h>
-#include <kernel/hashmap.h>
+# include <kernel/tokenize.h>
+# include <kernel/hashmap.h>
+#endif
 
 hashmap_t * kernel_args_map = NULL;
 
@@ -43,43 +63,94 @@ char * args_value(const char * karg) {
  *
  * @param arg A string containing all arguments, separated by spaces.
  */
-void args_parse(const char * _arg) {
+void args_parse(const char * cmdline) {
 	/* Sanity check... */
-	if (!_arg) { return; }
+	if (!cmdline) { return; }
+	if (!kernel_args_map) { kernel_args_map = hashmap_create(10); }
+	char * argbuf = strdup(cmdline);
+	char * x = argbuf;
 
-	char * arg = strdup(_arg);
+	for (;;) {
-	char * argv[1024];
+		while (*x && *x == ' ') x++; /* skip spaces */
-	int argc = tokenize(arg, " ", argv);
+		if (!*x) break;
 
-	/* New let's parse the tokens into the arguments list so we can index by key */
+		char * value = NULL;
-	if (!kernel_args_map) {
+		char * key = x;
-		kernel_args_map = hashmap_create(10);
+		while (*x && *x != '=' && *x != ' ') x++;
+
+		if (*x == '=') {
+			*x++ = '\0';
+			if (*x == '"') {
+				/* Start of quoted value */
+				x++;
+				value = x;
+				char * w = x;
+
+				while (*x && *x != '"') {
+					if (*x == '\\') {
+						x++;
+						switch (*x) {
+							case '"':  *w++ = '"';  x++; break;
+							case '\\': *w++ = '\\'; x++; break;
+							case '\0': goto _parse_error;
+							default:   *w++ = '\\'; *w++ = *x++; break;
+						}
+					} else {
+						*w++ = *x++;
+					}
 				}
 
-	for (int i = 0; i < argc; ++i) {
+				if (*x == '"') {
-		char * c = strdup(argv[i]);
+					*w = '\0';
-
+					x++;
-		char * name;
+				} else if (*x == '\0') {
-		char * value;
+					goto _parse_error;
-
-		name = c;
-		value = NULL;
-		/* Find the first = and replace it with a null */
-		char * v = c;
-		while (*v) {
-			if (*v == '=') {
-				*v = '\0';
-				v++;
-				value = v;
-				goto _break;
 				}
-			v++;
+			} else {
+				/* Start of unquoted value */
+				value = x;
+				while (*x && *x != ' ') x++;
+				if (*x == ' ') {
+					*x = '\0';
+					x++;
+				}
+			}
+		} else if (*x == ' ') {
+			/* Value-less argument where we need to set nil byte */
+			*x++ = '\0';
 		}
 
-_break:
+		hashmap_set(kernel_args_map, key, value ? strdup(value) : NULL);
-		hashmap_set(kernel_args_map, name, value);
 	}
 
-	free(arg);
+_parse_error:
+	free(argbuf);
+	return;
 }
 
+#ifndef _KERNEL_
+char * args_from_procfs(void) {
+	/* Open */
+	FILE * f = fopen("/proc/cmdline", "r");
+	if (!f) return NULL;
+
+	/* Determine size */
+	fseek(f, 0, SEEK_END);
+	size_t size = ftell(f);
+	fseek(f, 0, SEEK_SET);
+
+	/* Read */
+	char * cmdline = calloc(size + 1, 1);
+	size_t rsize = fread(cmdline, 1, size, f);
+	if (cmdline[rsize-1] == '\n') cmdline[rsize-1] = '\0';
+	fclose(f);
+
+	/* Parse */
+	kernel_args_map = hashmap_create(10);
+	args_parse(cmdline);
+
+	return cmdline;
+}
+
+#endif
+