blob: 7a7ae9e8d1a860439c83c079c29a7ca94599d4a5 [file] [log] [blame]
Anthony Liguori4a5fcab2009-11-11 10:39:23 -06001/*
Eric Blake6e8e5cb2016-01-29 06:48:37 -07002 * JSON Parser
Anthony Liguori4a5fcab2009-11-11 10:39:23 -06003 *
4 * Copyright IBM, Corp. 2009
5 *
6 * Authors:
7 * Anthony Liguori <aliguori@us.ibm.com>
8 *
9 * This work is licensed under the terms of the GNU LGPL, version 2.1 or later.
10 * See the COPYING.LIB file in the top-level directory.
11 *
12 */
13
Peter Maydellf2ad72b2016-01-29 17:50:01 +000014#include "qemu/osdep.h"
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +040015#include "qemu/cutils.h"
Markus Armbrustere59f39d2018-08-23 18:39:49 +020016#include "qemu/unicode.h"
Markus Armbrusterda34e652016-03-14 09:01:28 +010017#include "qapi/error.h"
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060018#include "qemu-common.h"
Markus Armbruster6b673952018-02-01 12:18:35 +010019#include "qapi/qmp/qbool.h"
Markus Armbruster452fcdb2018-02-01 12:18:39 +010020#include "qapi/qmp/qdict.h"
Markus Armbruster47e6b292018-02-01 12:18:38 +010021#include "qapi/qmp/qlist.h"
Markus Armbruster15280c32018-02-01 12:18:36 +010022#include "qapi/qmp/qnull.h"
23#include "qapi/qmp/qnum.h"
Markus Armbruster6b673952018-02-01 12:18:35 +010024#include "qapi/qmp/qstring.h"
Markus Armbruster86cdf9e2018-08-23 18:40:20 +020025#include "json-parser-int.h"
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060026
Markus Armbrusterabe7c202018-08-23 18:40:18 +020027struct JSONToken {
28 JSONTokenType type;
29 int x;
30 int y;
31 char str[];
32};
33
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060034typedef struct JSONParserContext
35{
Anthony Liguorief749d02011-06-01 12:14:50 -050036 Error *err;
Paolo Bonzini9bada892015-11-25 22:23:32 +010037 JSONToken *current;
Paolo Bonzini95385fe2015-11-25 22:23:31 +010038 GQueue *buf;
Markus Armbrusterada74c32018-08-23 18:40:22 +020039 va_list *ap;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060040} JSONParserContext;
41
42#define BUG_ON(cond) assert(!(cond))
43
44/**
45 * TODO
46 *
47 * 0) make errors meaningful again
48 * 1) add geometry information to tokens
49 * 3) should we return a parsed size?
50 * 4) deal with premature EOI
51 */
52
Markus Armbrusterada74c32018-08-23 18:40:22 +020053static QObject *parse_value(JSONParserContext *ctxt);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060054
55/**
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060056 * Error handler
57 */
Stefan Weil8b7968f2010-09-23 21:28:05 +020058static void GCC_FMT_ATTR(3, 4) parse_error(JSONParserContext *ctxt,
Paolo Bonzini9bada892015-11-25 22:23:32 +010059 JSONToken *token, const char *msg, ...)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060060{
Amos Kongc96c84a2010-03-24 23:12:05 +080061 va_list ap;
Anthony Liguorief749d02011-06-01 12:14:50 -050062 char message[1024];
Markus Armbruster574bf162018-08-23 18:39:50 +020063
64 if (ctxt->err) {
65 return;
66 }
Amos Kongc96c84a2010-03-24 23:12:05 +080067 va_start(ap, msg);
Anthony Liguorief749d02011-06-01 12:14:50 -050068 vsnprintf(message, sizeof(message), msg, ap);
Amos Kongc96c84a2010-03-24 23:12:05 +080069 va_end(ap);
Cole Robinsonf231b882014-03-21 19:42:26 -040070 error_setg(&ctxt->err, "JSON parse error, %s", message);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060071}
72
Markus Armbrusterdc45a072018-08-23 18:39:56 +020073static int cvt4hex(const char *s)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060074{
Markus Armbrusterdc45a072018-08-23 18:39:56 +020075 int cp, i;
76
77 cp = 0;
78 for (i = 0; i < 4; i++) {
79 if (!qemu_isxdigit(s[i])) {
80 return -1;
81 }
82 cp <<= 4;
83 if (s[i] >= '0' && s[i] <= '9') {
84 cp |= s[i] - '0';
85 } else if (s[i] >= 'a' && s[i] <= 'f') {
86 cp |= 10 + s[i] - 'a';
87 } else if (s[i] >= 'A' && s[i] <= 'F') {
88 cp |= 10 + s[i] - 'A';
89 } else {
90 return -1;
91 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060092 }
Markus Armbrusterdc45a072018-08-23 18:39:56 +020093 return cp;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060094}
95
96/**
Markus Armbrusterb2da4a42018-08-23 18:39:53 +020097 * parse_string(): Parse a JSON string
Anthony Liguori4a5fcab2009-11-11 10:39:23 -060098 *
Markus Armbrusterb2da4a42018-08-23 18:39:53 +020099 * From RFC 8259 "The JavaScript Object Notation (JSON) Data
100 * Interchange Format":
101 *
102 * char = unescaped /
103 * escape (
104 * %x22 / ; " quotation mark U+0022
105 * %x5C / ; \ reverse solidus U+005C
106 * %x2F / ; / solidus U+002F
107 * %x62 / ; b backspace U+0008
108 * %x66 / ; f form feed U+000C
109 * %x6E / ; n line feed U+000A
110 * %x72 / ; r carriage return U+000D
111 * %x74 / ; t tab U+0009
112 * %x75 4HEXDIG ) ; uXXXX U+XXXX
113 * escape = %x5C ; \
114 * quotation-mark = %x22 ; "
115 * unescaped = %x20-21 / %x23-5B / %x5D-10FFFF
116 *
117 * Extensions over RFC 8259:
118 * - Extra escape sequence in strings:
119 * 0x27 (apostrophe) is recognized after escape, too
120 * - Single-quoted strings:
121 * Like double-quoted strings, except they're delimited by %x27
122 * (apostrophe) instead of %x22 (quotation mark), and can't contain
123 * unescaped apostrophe, but can contain unescaped quotation mark.
124 *
125 * Note:
126 * - Encoding is modified UTF-8.
127 * - Invalid Unicode characters are rejected.
128 * - Control characters \x00..\x1F are rejected by the lexer.
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600129 */
Markus Armbrusterb2da4a42018-08-23 18:39:53 +0200130static QString *parse_string(JSONParserContext *ctxt, JSONToken *token)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600131{
Paolo Bonzini9bada892015-11-25 22:23:32 +0100132 const char *ptr = token->str;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600133 QString *str;
Markus Armbruster00ea57f2018-08-23 18:39:47 +0200134 char quote;
Markus Armbrusterdc45a072018-08-23 18:39:56 +0200135 const char *beg;
136 int cp, trailing;
Markus Armbrustere59f39d2018-08-23 18:39:49 +0200137 char *end;
138 ssize_t len;
139 char utf8_buf[5];
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600140
Markus Armbruster00ea57f2018-08-23 18:39:47 +0200141 assert(*ptr == '"' || *ptr == '\'');
142 quote = *ptr++;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600143 str = qstring_new();
Markus Armbruster00ea57f2018-08-23 18:39:47 +0200144
145 while (*ptr != quote) {
146 assert(*ptr);
Markus Armbruster16a48592018-08-23 18:40:23 +0200147 switch (*ptr) {
148 case '\\':
Markus Armbrusterdc45a072018-08-23 18:39:56 +0200149 beg = ptr++;
Markus Armbruster00ea57f2018-08-23 18:39:47 +0200150 switch (*ptr++) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600151 case '"':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200152 qstring_append_chr(str, '"');
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600153 break;
154 case '\'':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200155 qstring_append_chr(str, '\'');
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600156 break;
157 case '\\':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200158 qstring_append_chr(str, '\\');
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600159 break;
160 case '/':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200161 qstring_append_chr(str, '/');
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600162 break;
163 case 'b':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200164 qstring_append_chr(str, '\b');
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600165 break;
Luiz Capitulinobd032692010-05-19 17:06:15 -0300166 case 'f':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200167 qstring_append_chr(str, '\f');
Luiz Capitulinobd032692010-05-19 17:06:15 -0300168 break;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600169 case 'n':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200170 qstring_append_chr(str, '\n');
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600171 break;
172 case 'r':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200173 qstring_append_chr(str, '\r');
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600174 break;
175 case 't':
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200176 qstring_append_chr(str, '\t');
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600177 break;
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200178 case 'u':
Markus Armbrusterdc45a072018-08-23 18:39:56 +0200179 cp = cvt4hex(ptr);
180 ptr += 4;
181
182 /* handle surrogate pairs */
183 if (cp >= 0xD800 && cp <= 0xDBFF
184 && ptr[0] == '\\' && ptr[1] == 'u') {
185 /* leading surrogate followed by \u */
186 cp = 0x10000 + ((cp & 0x3FF) << 10);
187 trailing = cvt4hex(ptr + 2);
188 if (trailing >= 0xDC00 && trailing <= 0xDFFF) {
189 /* followed by trailing surrogate */
190 cp |= trailing & 0x3FF;
191 ptr += 6;
192 } else {
193 cp = -1; /* invalid */
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600194 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600195 }
196
Markus Armbruster46a628b2018-08-23 18:39:55 +0200197 if (mod_utf8_encode(utf8_buf, sizeof(utf8_buf), cp) < 0) {
198 parse_error(ctxt, token,
Markus Armbrusterdc45a072018-08-23 18:39:56 +0200199 "%.*s is not a valid Unicode character",
200 (int)(ptr - beg), beg);
Markus Armbruster46a628b2018-08-23 18:39:55 +0200201 goto out;
202 }
Markus Armbrusterde6decf2018-08-23 18:39:54 +0200203 qstring_append(str, utf8_buf);
204 break;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600205 default:
206 parse_error(ctxt, token, "invalid escape sequence in string");
207 goto out;
208 }
Markus Armbruster16a48592018-08-23 18:40:23 +0200209 break;
210 case '%':
Markus Armbruster8bca4612018-08-23 18:40:24 +0200211 if (ctxt->ap && ptr[1] != '%') {
Markus Armbruster16a48592018-08-23 18:40:23 +0200212 parse_error(ctxt, token, "can't interpolate into string");
213 goto out;
214 }
Markus Armbruster8bca4612018-08-23 18:40:24 +0200215 ptr++;
Markus Armbruster16a48592018-08-23 18:40:23 +0200216 /* fall through */
217 default:
Markus Armbrustere59f39d2018-08-23 18:39:49 +0200218 cp = mod_utf8_codepoint(ptr, 6, &end);
Markus Armbruster4b1c0cd2018-08-23 18:39:52 +0200219 if (cp < 0) {
Markus Armbrustere59f39d2018-08-23 18:39:49 +0200220 parse_error(ctxt, token, "invalid UTF-8 sequence in string");
221 goto out;
222 }
223 ptr = end;
224 len = mod_utf8_encode(utf8_buf, sizeof(utf8_buf), cp);
225 assert(len >= 0);
226 qstring_append(str, utf8_buf);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600227 }
228 }
229
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600230 return str;
231
232out:
Marc-André Lureaucb3e7f02018-04-19 17:01:43 +0200233 qobject_unref(str);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600234 return NULL;
235}
236
Paolo Bonzini9bada892015-11-25 22:23:32 +0100237/* Note: the token object returned by parser_context_peek_token or
238 * parser_context_pop_token is deleted as soon as parser_context_pop_token
239 * is called again.
Paolo Bonzini95385fe2015-11-25 22:23:31 +0100240 */
Paolo Bonzini9bada892015-11-25 22:23:32 +0100241static JSONToken *parser_context_pop_token(JSONParserContext *ctxt)
Michael Roth65c0f1e2012-08-15 13:45:43 -0500242{
Paolo Bonzini9bada892015-11-25 22:23:32 +0100243 g_free(ctxt->current);
Paolo Bonzini95385fe2015-11-25 22:23:31 +0100244 ctxt->current = g_queue_pop_head(ctxt->buf);
245 return ctxt->current;
Michael Roth65c0f1e2012-08-15 13:45:43 -0500246}
247
Paolo Bonzini9bada892015-11-25 22:23:32 +0100248static JSONToken *parser_context_peek_token(JSONParserContext *ctxt)
Michael Roth65c0f1e2012-08-15 13:45:43 -0500249{
Paolo Bonzini95385fe2015-11-25 22:23:31 +0100250 return g_queue_peek_head(ctxt->buf);
Michael Roth65c0f1e2012-08-15 13:45:43 -0500251}
252
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600253/**
254 * Parsing rules
255 */
Markus Armbrusterada74c32018-08-23 18:40:22 +0200256static int parse_pair(JSONParserContext *ctxt, QDict *dict)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600257{
Max Reitz532fb532018-03-10 16:14:36 -0600258 QObject *value;
259 QString *key = NULL;
Paolo Bonzini9bada892015-11-25 22:23:32 +0100260 JSONToken *peek, *token;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600261
Michael Roth65c0f1e2012-08-15 13:45:43 -0500262 peek = parser_context_peek_token(ctxt);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500263 if (peek == NULL) {
264 parse_error(ctxt, NULL, "premature EOI");
265 goto out;
266 }
267
Markus Armbrusterada74c32018-08-23 18:40:22 +0200268 key = qobject_to(QString, parse_value(ctxt));
Max Reitz532fb532018-03-10 16:14:36 -0600269 if (!key) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600270 parse_error(ctxt, peek, "key is not a string in object");
271 goto out;
272 }
273
Michael Roth65c0f1e2012-08-15 13:45:43 -0500274 token = parser_context_pop_token(ctxt);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500275 if (token == NULL) {
276 parse_error(ctxt, NULL, "premature EOI");
277 goto out;
278 }
279
Paolo Bonzini9bada892015-11-25 22:23:32 +0100280 if (token->type != JSON_COLON) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600281 parse_error(ctxt, token, "missing : in object pair");
282 goto out;
283 }
284
Markus Armbrusterada74c32018-08-23 18:40:22 +0200285 value = parse_value(ctxt);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600286 if (value == NULL) {
287 parse_error(ctxt, token, "Missing value in dict");
288 goto out;
289 }
290
Markus Armbruster00382fa2018-12-06 13:17:43 +0100291 if (qdict_haskey(dict, qstring_get_str(key))) {
292 parse_error(ctxt, token, "duplicate key");
293 goto out;
294 }
295
Max Reitz532fb532018-03-10 16:14:36 -0600296 qdict_put_obj(dict, qstring_get_str(key), value);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600297
Marc-André Lureaucb3e7f02018-04-19 17:01:43 +0200298 qobject_unref(key);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600299
300 return 0;
301
302out:
Marc-André Lureaucb3e7f02018-04-19 17:01:43 +0200303 qobject_unref(key);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600304
305 return -1;
306}
307
Markus Armbrusterada74c32018-08-23 18:40:22 +0200308static QObject *parse_object(JSONParserContext *ctxt)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600309{
310 QDict *dict = NULL;
Paolo Bonzini9bada892015-11-25 22:23:32 +0100311 JSONToken *token, *peek;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600312
Michael Roth65c0f1e2012-08-15 13:45:43 -0500313 token = parser_context_pop_token(ctxt);
Paolo Bonzini9bada892015-11-25 22:23:32 +0100314 assert(token && token->type == JSON_LCURLY);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600315
316 dict = qdict_new();
317
Michael Roth65c0f1e2012-08-15 13:45:43 -0500318 peek = parser_context_peek_token(ctxt);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500319 if (peek == NULL) {
320 parse_error(ctxt, NULL, "premature EOI");
321 goto out;
322 }
323
Paolo Bonzini9bada892015-11-25 22:23:32 +0100324 if (peek->type != JSON_RCURLY) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200325 if (parse_pair(ctxt, dict) == -1) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600326 goto out;
327 }
328
Michael Roth65c0f1e2012-08-15 13:45:43 -0500329 token = parser_context_pop_token(ctxt);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500330 if (token == NULL) {
331 parse_error(ctxt, NULL, "premature EOI");
332 goto out;
333 }
334
Paolo Bonzini9bada892015-11-25 22:23:32 +0100335 while (token->type != JSON_RCURLY) {
336 if (token->type != JSON_COMMA) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600337 parse_error(ctxt, token, "expected separator in dict");
338 goto out;
339 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600340
Markus Armbrusterada74c32018-08-23 18:40:22 +0200341 if (parse_pair(ctxt, dict) == -1) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600342 goto out;
343 }
344
Michael Roth65c0f1e2012-08-15 13:45:43 -0500345 token = parser_context_pop_token(ctxt);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500346 if (token == NULL) {
347 parse_error(ctxt, NULL, "premature EOI");
348 goto out;
349 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600350 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600351 } else {
Gongleia491af42014-06-10 17:20:24 +0800352 (void)parser_context_pop_token(ctxt);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600353 }
354
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600355 return QOBJECT(dict);
356
357out:
Marc-André Lureaucb3e7f02018-04-19 17:01:43 +0200358 qobject_unref(dict);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600359 return NULL;
360}
361
Markus Armbrusterada74c32018-08-23 18:40:22 +0200362static QObject *parse_array(JSONParserContext *ctxt)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600363{
364 QList *list = NULL;
Paolo Bonzini9bada892015-11-25 22:23:32 +0100365 JSONToken *token, *peek;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600366
Michael Roth65c0f1e2012-08-15 13:45:43 -0500367 token = parser_context_pop_token(ctxt);
Paolo Bonzini9bada892015-11-25 22:23:32 +0100368 assert(token && token->type == JSON_LSQUARE);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600369
370 list = qlist_new();
371
Michael Roth65c0f1e2012-08-15 13:45:43 -0500372 peek = parser_context_peek_token(ctxt);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500373 if (peek == NULL) {
374 parse_error(ctxt, NULL, "premature EOI");
375 goto out;
376 }
377
Paolo Bonzini9bada892015-11-25 22:23:32 +0100378 if (peek->type != JSON_RSQUARE) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600379 QObject *obj;
380
Markus Armbrusterada74c32018-08-23 18:40:22 +0200381 obj = parse_value(ctxt);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600382 if (obj == NULL) {
383 parse_error(ctxt, token, "expecting value");
384 goto out;
385 }
386
387 qlist_append_obj(list, obj);
388
Michael Roth65c0f1e2012-08-15 13:45:43 -0500389 token = parser_context_pop_token(ctxt);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500390 if (token == NULL) {
391 parse_error(ctxt, NULL, "premature EOI");
392 goto out;
393 }
394
Paolo Bonzini9bada892015-11-25 22:23:32 +0100395 while (token->type != JSON_RSQUARE) {
396 if (token->type != JSON_COMMA) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600397 parse_error(ctxt, token, "expected separator in list");
398 goto out;
399 }
400
Markus Armbrusterada74c32018-08-23 18:40:22 +0200401 obj = parse_value(ctxt);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600402 if (obj == NULL) {
403 parse_error(ctxt, token, "expecting value");
404 goto out;
405 }
406
407 qlist_append_obj(list, obj);
408
Michael Roth65c0f1e2012-08-15 13:45:43 -0500409 token = parser_context_pop_token(ctxt);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500410 if (token == NULL) {
411 parse_error(ctxt, NULL, "premature EOI");
412 goto out;
413 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600414 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600415 } else {
Gongleia491af42014-06-10 17:20:24 +0800416 (void)parser_context_pop_token(ctxt);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600417 }
418
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600419 return QOBJECT(list);
420
421out:
Marc-André Lureaucb3e7f02018-04-19 17:01:43 +0200422 qobject_unref(list);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600423 return NULL;
424}
425
Michael Roth65c0f1e2012-08-15 13:45:43 -0500426static QObject *parse_keyword(JSONParserContext *ctxt)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600427{
Paolo Bonzini9bada892015-11-25 22:23:32 +0100428 JSONToken *token;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600429
Michael Roth65c0f1e2012-08-15 13:45:43 -0500430 token = parser_context_pop_token(ctxt);
Paolo Bonzini9bada892015-11-25 22:23:32 +0100431 assert(token && token->type == JSON_KEYWORD);
Markus Armbruster50e2a462015-11-25 22:23:27 +0100432
Paolo Bonzini9bada892015-11-25 22:23:32 +0100433 if (!strcmp(token->str, "true")) {
Markus Armbrusterd538b252015-11-25 22:23:30 +0100434 return QOBJECT(qbool_from_bool(true));
Paolo Bonzini9bada892015-11-25 22:23:32 +0100435 } else if (!strcmp(token->str, "false")) {
Markus Armbrusterd538b252015-11-25 22:23:30 +0100436 return QOBJECT(qbool_from_bool(false));
Paolo Bonzini9bada892015-11-25 22:23:32 +0100437 } else if (!strcmp(token->str, "null")) {
Markus Armbruster006ca092017-06-26 13:52:24 +0200438 return QOBJECT(qnull());
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600439 }
Paolo Bonzini9bada892015-11-25 22:23:32 +0100440 parse_error(ctxt, token, "invalid keyword '%s'", token->str);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600441 return NULL;
442}
443
Markus Armbrusterada74c32018-08-23 18:40:22 +0200444static QObject *parse_interpolation(JSONParserContext *ctxt)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600445{
Paolo Bonzini9bada892015-11-25 22:23:32 +0100446 JSONToken *token;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600447
Michael Roth65c0f1e2012-08-15 13:45:43 -0500448 token = parser_context_pop_token(ctxt);
Markus Armbruster61030282018-08-23 18:40:04 +0200449 assert(token && token->type == JSON_INTERP);
Markus Armbruster6b9606f2015-11-25 22:23:28 +0100450
Paolo Bonzini9bada892015-11-25 22:23:32 +0100451 if (!strcmp(token->str, "%p")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200452 return va_arg(*ctxt->ap, QObject *);
Paolo Bonzini9bada892015-11-25 22:23:32 +0100453 } else if (!strcmp(token->str, "%i")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200454 return QOBJECT(qbool_from_bool(va_arg(*ctxt->ap, int)));
Paolo Bonzini9bada892015-11-25 22:23:32 +0100455 } else if (!strcmp(token->str, "%d")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200456 return QOBJECT(qnum_from_int(va_arg(*ctxt->ap, int)));
Paolo Bonzini9bada892015-11-25 22:23:32 +0100457 } else if (!strcmp(token->str, "%ld")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200458 return QOBJECT(qnum_from_int(va_arg(*ctxt->ap, long)));
Markus Armbruster53a0d612018-08-23 18:40:08 +0200459 } else if (!strcmp(token->str, "%lld")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200460 return QOBJECT(qnum_from_int(va_arg(*ctxt->ap, long long)));
Markus Armbruster53a0d612018-08-23 18:40:08 +0200461 } else if (!strcmp(token->str, "%" PRId64)) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200462 return QOBJECT(qnum_from_int(va_arg(*ctxt->ap, int64_t)));
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +0400463 } else if (!strcmp(token->str, "%u")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200464 return QOBJECT(qnum_from_uint(va_arg(*ctxt->ap, unsigned int)));
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +0400465 } else if (!strcmp(token->str, "%lu")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200466 return QOBJECT(qnum_from_uint(va_arg(*ctxt->ap, unsigned long)));
Markus Armbruster53a0d612018-08-23 18:40:08 +0200467 } else if (!strcmp(token->str, "%llu")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200468 return QOBJECT(qnum_from_uint(va_arg(*ctxt->ap, unsigned long long)));
Markus Armbruster53a0d612018-08-23 18:40:08 +0200469 } else if (!strcmp(token->str, "%" PRIu64)) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200470 return QOBJECT(qnum_from_uint(va_arg(*ctxt->ap, uint64_t)));
Paolo Bonzini9bada892015-11-25 22:23:32 +0100471 } else if (!strcmp(token->str, "%s")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200472 return QOBJECT(qstring_from_str(va_arg(*ctxt->ap, const char *)));
Paolo Bonzini9bada892015-11-25 22:23:32 +0100473 } else if (!strcmp(token->str, "%f")) {
Markus Armbrusterada74c32018-08-23 18:40:22 +0200474 return QOBJECT(qnum_from_double(va_arg(*ctxt->ap, double)));
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600475 }
Markus Armbrusterf7617d42018-08-23 18:40:07 +0200476 parse_error(ctxt, token, "invalid interpolation '%s'", token->str);
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600477 return NULL;
478}
479
Michael Roth65c0f1e2012-08-15 13:45:43 -0500480static QObject *parse_literal(JSONParserContext *ctxt)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600481{
Paolo Bonzini9bada892015-11-25 22:23:32 +0100482 JSONToken *token;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600483
Michael Roth65c0f1e2012-08-15 13:45:43 -0500484 token = parser_context_pop_token(ctxt);
Markus Armbrusterd538b252015-11-25 22:23:30 +0100485 assert(token);
Anthony Liguori11e8a462011-06-01 12:14:55 -0500486
Paolo Bonzini9bada892015-11-25 22:23:32 +0100487 switch (token->type) {
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600488 case JSON_STRING:
Markus Armbrusterb2da4a42018-08-23 18:39:53 +0200489 return QOBJECT(parse_string(ctxt, token));
Michael Roth3d5b3ec2013-05-10 17:46:05 -0500490 case JSON_INTEGER: {
Marc-André Lureau01b2ffc2017-06-07 20:35:58 +0400491 /*
492 * Represent JSON_INTEGER as QNUM_I64 if possible, else as
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +0400493 * QNUM_U64, else as QNUM_DOUBLE. Note that qemu_strtoi64()
494 * and qemu_strtou64() fail with ERANGE when it's not
495 * possible.
Michael Roth3d5b3ec2013-05-10 17:46:05 -0500496 *
Marc-André Lureau01b2ffc2017-06-07 20:35:58 +0400497 * qnum_get_int() will then work for any signed 64-bit
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +0400498 * JSON_INTEGER, qnum_get_uint() for any unsigned 64-bit
499 * integer, and qnum_get_double() both for any JSON_INTEGER
500 * and any JSON_FLOAT (with precision loss for integers beyond
501 * 53 bits)
Michael Roth3d5b3ec2013-05-10 17:46:05 -0500502 */
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +0400503 int ret;
Michael Roth3d5b3ec2013-05-10 17:46:05 -0500504 int64_t value;
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +0400505 uint64_t uvalue;
Michael Roth3d5b3ec2013-05-10 17:46:05 -0500506
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +0400507 ret = qemu_strtoi64(token->str, NULL, 10, &value);
508 if (!ret) {
Marc-André Lureau01b2ffc2017-06-07 20:35:58 +0400509 return QOBJECT(qnum_from_int(value));
Michael Roth3d5b3ec2013-05-10 17:46:05 -0500510 }
Marc-André Lureau2bc7cfe2017-06-07 20:36:02 +0400511 assert(ret == -ERANGE);
512
513 if (token->str[0] != '-') {
514 ret = qemu_strtou64(token->str, NULL, 10, &uvalue);
515 if (!ret) {
516 return QOBJECT(qnum_from_uint(uvalue));
517 }
518 assert(ret == -ERANGE);
519 }
Michael Roth3d5b3ec2013-05-10 17:46:05 -0500520 /* fall through to JSON_FLOAT */
521 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600522 case JSON_FLOAT:
Eric Blake6e8e5cb2016-01-29 06:48:37 -0700523 /* FIXME dependent on locale; a pervasive issue in QEMU */
Markus Armbruster37aded92018-08-23 18:40:25 +0200524 /* FIXME our lexer matches RFC 8259 in forbidding Inf or NaN,
Eric Blake6e8e5cb2016-01-29 06:48:37 -0700525 * but those might be useful extensions beyond JSON */
Marc-André Lureau01b2ffc2017-06-07 20:35:58 +0400526 return QOBJECT(qnum_from_double(strtod(token->str, NULL)));
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600527 default:
Markus Armbrusterd538b252015-11-25 22:23:30 +0100528 abort();
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600529 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600530}
531
Markus Armbrusterada74c32018-08-23 18:40:22 +0200532static QObject *parse_value(JSONParserContext *ctxt)
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600533{
Paolo Bonzini9bada892015-11-25 22:23:32 +0100534 JSONToken *token;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600535
Markus Armbrusterd538b252015-11-25 22:23:30 +0100536 token = parser_context_peek_token(ctxt);
537 if (token == NULL) {
538 parse_error(ctxt, NULL, "premature EOI");
539 return NULL;
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600540 }
541
Paolo Bonzini9bada892015-11-25 22:23:32 +0100542 switch (token->type) {
Markus Armbrusterd538b252015-11-25 22:23:30 +0100543 case JSON_LCURLY:
Markus Armbrusterada74c32018-08-23 18:40:22 +0200544 return parse_object(ctxt);
Markus Armbrusterd538b252015-11-25 22:23:30 +0100545 case JSON_LSQUARE:
Markus Armbrusterada74c32018-08-23 18:40:22 +0200546 return parse_array(ctxt);
Markus Armbruster61030282018-08-23 18:40:04 +0200547 case JSON_INTERP:
Markus Armbrusterada74c32018-08-23 18:40:22 +0200548 return parse_interpolation(ctxt);
Markus Armbrusterd538b252015-11-25 22:23:30 +0100549 case JSON_INTEGER:
550 case JSON_FLOAT:
551 case JSON_STRING:
552 return parse_literal(ctxt);
553 case JSON_KEYWORD:
554 return parse_keyword(ctxt);
555 default:
556 parse_error(ctxt, token, "expecting value");
557 return NULL;
558 }
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600559}
560
Markus Armbrusterabe7c202018-08-23 18:40:18 +0200561JSONToken *json_token(JSONTokenType type, int x, int y, GString *tokstr)
562{
563 JSONToken *token = g_malloc(sizeof(JSONToken) + tokstr->len + 1);
564
565 token->type = type;
566 memcpy(token->str, tokstr->str, tokstr->len);
567 token->str[tokstr->len] = 0;
568 token->x = x;
569 token->y = y;
570 return token;
571}
572
Markus Armbruster62815d82018-08-23 18:40:01 +0200573QObject *json_parser_parse(GQueue *tokens, va_list *ap, Error **errp)
Anthony Liguorief749d02011-06-01 12:14:50 -0500574{
Markus Armbrusterada74c32018-08-23 18:40:22 +0200575 JSONParserContext ctxt = { .buf = tokens, .ap = ap };
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600576 QObject *result;
577
Markus Armbrusterada74c32018-08-23 18:40:22 +0200578 result = parse_value(&ctxt);
Markus Armbruster5d501132018-08-23 18:40:13 +0200579 assert(ctxt.err || g_queue_is_empty(ctxt.buf));
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600580
Marc-André Lureaue8b19d72018-08-23 18:39:59 +0200581 error_propagate(errp, ctxt.err);
Michael Roth65c0f1e2012-08-15 13:45:43 -0500582
Marc-André Lureaue8b19d72018-08-23 18:39:59 +0200583 while (!g_queue_is_empty(ctxt.buf)) {
584 parser_context_pop_token(&ctxt);
585 }
586 g_free(ctxt.current);
Anthony Liguorief749d02011-06-01 12:14:50 -0500587
Anthony Liguori4a5fcab2009-11-11 10:39:23 -0600588 return result;
589}