xref: /linux/fs/unicode/utf8-selftest.c (revision af873fcecef567abf8a3468b06dd4e4aab46da6d)
1 /*
2  * Kernel module for testing utf-8 support.
3  *
4  * Copyright 2017 Collabora Ltd.
5  *
6  * This software is licensed under the terms of the GNU General Public
7  * License version 2, as published by the Free Software Foundation, and
8  * may be copied, distributed, and modified under those terms.
9  *
10  * This program is distributed in the hope that it will be useful,
11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13  * GNU General Public License for more details.
14  */
15 
16 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
17 
18 #include <linux/module.h>
19 #include <linux/printk.h>
20 #include <linux/unicode.h>
21 #include <linux/dcache.h>
22 
23 #include "utf8n.h"
24 
25 unsigned int failed_tests;
26 unsigned int total_tests;
27 
28 /* Tests will be based on this version. */
29 #define latest_maj 12
30 #define latest_min 1
31 #define latest_rev 0
32 
33 #define _test(cond, func, line, fmt, ...) do {				\
34 		total_tests++;						\
35 		if (!cond) {						\
36 			failed_tests++;					\
37 			pr_err("test %s:%d Failed: %s%s",		\
38 			       func, line, #cond, (fmt?":":"."));	\
39 			if (fmt)					\
40 				pr_err(fmt, ##__VA_ARGS__);		\
41 		}							\
42 	} while (0)
43 #define test_f(cond, fmt, ...) _test(cond, __func__, __LINE__, fmt, ##__VA_ARGS__)
44 #define test(cond) _test(cond, __func__, __LINE__, "")
45 
46 const static struct {
47 	/* UTF-8 strings in this vector _must_ be NULL-terminated. */
48 	unsigned char str[10];
49 	unsigned char dec[10];
50 } nfdi_test_data[] = {
51 	/* Trivial sequence */
52 	{
53 		/* "ABba" decomposes to itself */
54 		.str = "aBba",
55 		.dec = "aBba",
56 	},
57 	/* Simple equivalent sequences */
58 	{
59                /* 'VULGAR FRACTION ONE QUARTER' cannot decompose to
60                   'NUMBER 1' + 'FRACTION SLASH' + 'NUMBER 4' on
61                   canonical decomposition */
62                .str = {0xc2, 0xbc, 0x00},
63 	       .dec = {0xc2, 0xbc, 0x00},
64 	},
65 	{
66 		/* 'LATIN SMALL LETTER A WITH DIAERESIS' decomposes to
67 		   'LETTER A' + 'COMBINING DIAERESIS' */
68 		.str = {0xc3, 0xa4, 0x00},
69 		.dec = {0x61, 0xcc, 0x88, 0x00},
70 	},
71 	{
72 		/* 'LATIN SMALL LETTER LJ' can't decompose to
73 		   'LETTER L' + 'LETTER J' on canonical decomposition */
74 		.str = {0xC7, 0x89, 0x00},
75 		.dec = {0xC7, 0x89, 0x00},
76 	},
77 	{
78 		/* GREEK ANO TELEIA decomposes to MIDDLE DOT */
79 		.str = {0xCE, 0x87, 0x00},
80 		.dec = {0xC2, 0xB7, 0x00}
81 	},
82 	/* Canonical ordering */
83 	{
84 		/* A + 'COMBINING ACUTE ACCENT' + 'COMBINING OGONEK' decomposes
85 		   to A + 'COMBINING OGONEK' + 'COMBINING ACUTE ACCENT' */
86 		.str = {0x41, 0xcc, 0x81, 0xcc, 0xa8, 0x0},
87 		.dec = {0x41, 0xcc, 0xa8, 0xcc, 0x81, 0x0},
88 	},
89 	{
90 		/* 'LATIN SMALL LETTER A WITH DIAERESIS' + 'COMBINING OGONEK'
91 		   decomposes to
92 		   'LETTER A' + 'COMBINING OGONEK' + 'COMBINING DIAERESIS' */
93 		.str = {0xc3, 0xa4, 0xCC, 0xA8, 0x00},
94 
95 		.dec = {0x61, 0xCC, 0xA8, 0xcc, 0x88, 0x00},
96 	},
97 
98 };
99 
100 const static struct {
101 	/* UTF-8 strings in this vector _must_ be NULL-terminated. */
102 	unsigned char str[30];
103 	unsigned char ncf[30];
104 } nfdicf_test_data[] = {
105 	/* Trivial sequences */
106 	{
107 		/* "ABba" folds to lowercase */
108 		.str = {0x41, 0x42, 0x62, 0x61, 0x00},
109 		.ncf = {0x61, 0x62, 0x62, 0x61, 0x00},
110 	},
111 	{
112 		/* All ASCII folds to lower-case */
113 		.str = "ABCDEFGHIJKLMNOPQRSTUVWXYZ0.1",
114 		.ncf = "abcdefghijklmnopqrstuvwxyz0.1",
115 	},
116 	{
117 		/* LATIN SMALL LETTER SHARP S folds to
118 		   LATIN SMALL LETTER S + LATIN SMALL LETTER S */
119 		.str = {0xc3, 0x9f, 0x00},
120 		.ncf = {0x73, 0x73, 0x00},
121 	},
122 	{
123 		/* LATIN CAPITAL LETTER A WITH RING ABOVE folds to
124 		   LATIN SMALL LETTER A + COMBINING RING ABOVE */
125 		.str = {0xC3, 0x85, 0x00},
126 		.ncf = {0x61, 0xcc, 0x8a, 0x00},
127 	},
128 	/* Introduced by UTF-8.0.0. */
129 	/* Cherokee letters are interesting test-cases because they fold
130 	   to upper-case.  Before 8.0.0, Cherokee lowercase were
131 	   undefined, thus, the folding from LC is not stable between
132 	   7.0.0 -> 8.0.0, but it is from UC. */
133 	{
134 		/* CHEROKEE SMALL LETTER A folds to CHEROKEE LETTER A */
135 		.str = {0xea, 0xad, 0xb0, 0x00},
136 		.ncf = {0xe1, 0x8e, 0xa0, 0x00},
137 	},
138 	{
139 		/* CHEROKEE SMALL LETTER YE folds to CHEROKEE LETTER YE */
140 		.str = {0xe1, 0x8f, 0xb8, 0x00},
141 		.ncf = {0xe1, 0x8f, 0xb0, 0x00},
142 	},
143 	{
144 		/* OLD HUNGARIAN CAPITAL LETTER AMB folds to
145 		   OLD HUNGARIAN SMALL LETTER AMB */
146 		.str = {0xf0, 0x90, 0xb2, 0x83, 0x00},
147 		.ncf = {0xf0, 0x90, 0xb3, 0x83, 0x00},
148 	},
149 	/* Introduced by UTF-9.0.0. */
150 	{
151 		/* OSAGE CAPITAL LETTER CHA folds to
152 		   OSAGE SMALL LETTER CHA */
153 		.str = {0xf0, 0x90, 0x92, 0xb5, 0x00},
154 		.ncf = {0xf0, 0x90, 0x93, 0x9d, 0x00},
155 	},
156 	{
157 		/* LATIN CAPITAL LETTER SMALL CAPITAL I folds to
158 		   LATIN LETTER SMALL CAPITAL I */
159 		.str = {0xea, 0x9e, 0xae, 0x00},
160 		.ncf = {0xc9, 0xaa, 0x00},
161 	},
162 	/* Introduced by UTF-11.0.0. */
163 	{
164 		/* GEORGIAN SMALL LETTER AN folds to GEORGIAN MTAVRULI
165 		   CAPITAL LETTER AN */
166 		.str = {0xe1, 0xb2, 0x90, 0x00},
167 		.ncf = {0xe1, 0x83, 0x90, 0x00},
168 	}
169 };
170 
171 static void check_utf8_nfdi(void)
172 {
173 	int i;
174 	struct utf8cursor u8c;
175 	const struct utf8data *data;
176 
177 	data = utf8nfdi(UNICODE_AGE(latest_maj, latest_min, latest_rev));
178 	if (!data) {
179 		pr_err("%s: Unable to load utf8-%d.%d.%d. Skipping.\n",
180 		       __func__, latest_maj, latest_min, latest_rev);
181 		return;
182 	}
183 
184 	for (i = 0; i < ARRAY_SIZE(nfdi_test_data); i++) {
185 		int len = strlen(nfdi_test_data[i].str);
186 		int nlen = strlen(nfdi_test_data[i].dec);
187 		int j = 0;
188 		unsigned char c;
189 
190 		test((utf8len(data, nfdi_test_data[i].str) == nlen));
191 		test((utf8nlen(data, nfdi_test_data[i].str, len) == nlen));
192 
193 		if (utf8cursor(&u8c, data, nfdi_test_data[i].str) < 0)
194 			pr_err("can't create cursor\n");
195 
196 		while ((c = utf8byte(&u8c)) > 0) {
197 			test_f((c == nfdi_test_data[i].dec[j]),
198 			       "Unexpected byte 0x%x should be 0x%x\n",
199 			       c, nfdi_test_data[i].dec[j]);
200 			j++;
201 		}
202 
203 		test((j == nlen));
204 	}
205 }
206 
207 static void check_utf8_nfdicf(void)
208 {
209 	int i;
210 	struct utf8cursor u8c;
211 	const struct utf8data *data;
212 
213 	data = utf8nfdicf(UNICODE_AGE(latest_maj, latest_min, latest_rev));
214 	if (!data) {
215 		pr_err("%s: Unable to load utf8-%d.%d.%d. Skipping.\n",
216 		       __func__, latest_maj, latest_min, latest_rev);
217 		return;
218 	}
219 
220 	for (i = 0; i < ARRAY_SIZE(nfdicf_test_data); i++) {
221 		int len = strlen(nfdicf_test_data[i].str);
222 		int nlen = strlen(nfdicf_test_data[i].ncf);
223 		int j = 0;
224 		unsigned char c;
225 
226 		test((utf8len(data, nfdicf_test_data[i].str) == nlen));
227 		test((utf8nlen(data, nfdicf_test_data[i].str, len) == nlen));
228 
229 		if (utf8cursor(&u8c, data, nfdicf_test_data[i].str) < 0)
230 			pr_err("can't create cursor\n");
231 
232 		while ((c = utf8byte(&u8c)) > 0) {
233 			test_f((c == nfdicf_test_data[i].ncf[j]),
234 			       "Unexpected byte 0x%x should be 0x%x\n",
235 			       c, nfdicf_test_data[i].ncf[j]);
236 			j++;
237 		}
238 
239 		test((j == nlen));
240 	}
241 }
242 
243 static void check_utf8_comparisons(void)
244 {
245 	int i;
246 	struct unicode_map *table = utf8_load("12.1.0");
247 
248 	if (IS_ERR(table)) {
249 		pr_err("%s: Unable to load utf8 %d.%d.%d. Skipping.\n",
250 		       __func__, latest_maj, latest_min, latest_rev);
251 		return;
252 	}
253 
254 	for (i = 0; i < ARRAY_SIZE(nfdi_test_data); i++) {
255 		const struct qstr s1 = {.name = nfdi_test_data[i].str,
256 					.len = sizeof(nfdi_test_data[i].str)};
257 		const struct qstr s2 = {.name = nfdi_test_data[i].dec,
258 					.len = sizeof(nfdi_test_data[i].dec)};
259 
260 		test_f(!utf8_strncmp(table, &s1, &s2),
261 		       "%s %s comparison mismatch\n", s1.name, s2.name);
262 	}
263 
264 	for (i = 0; i < ARRAY_SIZE(nfdicf_test_data); i++) {
265 		const struct qstr s1 = {.name = nfdicf_test_data[i].str,
266 					.len = sizeof(nfdicf_test_data[i].str)};
267 		const struct qstr s2 = {.name = nfdicf_test_data[i].ncf,
268 					.len = sizeof(nfdicf_test_data[i].ncf)};
269 
270 		test_f(!utf8_strncasecmp(table, &s1, &s2),
271 		       "%s %s comparison mismatch\n", s1.name, s2.name);
272 	}
273 
274 	utf8_unload(table);
275 }
276 
277 static void check_supported_versions(void)
278 {
279 	/* Unicode 7.0.0 should be supported. */
280 	test(utf8version_is_supported(7, 0, 0));
281 
282 	/* Unicode 9.0.0 should be supported. */
283 	test(utf8version_is_supported(9, 0, 0));
284 
285 	/* Unicode 1x.0.0 (the latest version) should be supported. */
286 	test(utf8version_is_supported(latest_maj, latest_min, latest_rev));
287 
288 	/* Next versions don't exist. */
289 	test(!utf8version_is_supported(13, 0, 0));
290 	test(!utf8version_is_supported(0, 0, 0));
291 	test(!utf8version_is_supported(-1, -1, -1));
292 }
293 
294 static int __init init_test_ucd(void)
295 {
296 	failed_tests = 0;
297 	total_tests = 0;
298 
299 	check_supported_versions();
300 	check_utf8_nfdi();
301 	check_utf8_nfdicf();
302 	check_utf8_comparisons();
303 
304 	if (!failed_tests)
305 		pr_info("All %u tests passed\n", total_tests);
306 	else
307 		pr_err("%u out of %u tests failed\n", failed_tests,
308 		       total_tests);
309 	return 0;
310 }
311 
312 static void __exit exit_test_ucd(void)
313 {
314 }
315 
316 module_init(init_test_ucd);
317 module_exit(exit_test_ucd);
318 
319 MODULE_AUTHOR("Gabriel Krisman Bertazi <krisman@collabora.co.uk>");
320 MODULE_LICENSE("GPL");
321