Resguardo inicial

linux-7.2.3/fs/unicode/.gitignore

@@ -0,0 +1,3 @@
+# SPDX-License-Identifier: GPL-2.0-only
+/mkutf8data
+/utf8data.c

linux-7.2.3/fs/unicode/Kconfig

@@ -0,0 +1,16 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# UTF-8 normalization
+#
+config UNICODE
+	tristate "UTF-8 normalization and casefolding support"
+	help
+	  Say Y here to enable UTF-8 NFD normalization and NFD+CF casefolding
+	  support.  If you say M here the large table of case foldings will
+	  be a separate loadable module that gets requested only when a file
+	  system actually use it.
+
+config UNICODE_NORMALIZATION_KUNIT_TEST
+	tristate "Test UTF-8 normalization support"
+	depends on UNICODE && KUNIT
+	default KUNIT_ALL_TESTS

linux-7.2.3/fs/unicode/Makefile

@@ -0,0 +1,41 @@
+# SPDX-License-Identifier: GPL-2.0
+
+ifneq ($(CONFIG_UNICODE),)
+obj-y			+= unicode.o
+endif
+obj-$(CONFIG_UNICODE)	+= utf8data.o
+obj-$(CONFIG_UNICODE_NORMALIZATION_KUNIT_TEST) += tests/utf8_kunit.o
+
+unicode-y := utf8-norm.o utf8-core.o
+
+$(obj)/utf8-data.o: $(obj)/utf8data.c
+
+# In the normal build, the checked-in utf8data.c is just shipped.
+#
+# To generate utf8data.c from UCD, put *.txt files in this directory
+# and pass REGENERATE_UTF8DATA=1 from the command line.
+ifdef REGENERATE_UTF8DATA
+
+quiet_cmd_utf8data = GEN     $@
+      cmd_utf8data = $< \
+		-a $(src)/DerivedAge.txt \
+		-c $(src)/DerivedCombiningClass.txt \
+		-p $(src)/DerivedCoreProperties.txt \
+		-d $(src)/UnicodeData.txt \
+		-f $(src)/CaseFolding.txt \
+		-n $(src)/NormalizationCorrections.txt \
+		-t $(src)/NormalizationTest.txt \
+		-o $@
+
+$(obj)/utf8data.c: $(obj)/mkutf8data $(filter %.txt, $(cmd_utf8data)) FORCE
+	$(call if_changed,utf8data)
+
+else
+
+$(obj)/utf8data.c: $(src)/utf8data.c_shipped FORCE
+	$(call if_changed,copy)
+
+endif
+
+targets += utf8data.c
+hostprogs += mkutf8data

linux-7.2.3/fs/unicode/README.utf8data

@@ -0,0 +1,57 @@
+The utf8data.c file in this directory is generated from the Unicode
+Character Database for version 12.1.0 of the Unicode standard.
+
+The full set of files can be found here:
+
+  http://www.unicode.org/Public/12.1.0/ucd/
+
+Individual source links:
+
+  https://www.unicode.org/Public/12.1.0/ucd/CaseFolding.txt
+  https://www.unicode.org/Public/12.1.0/ucd/DerivedAge.txt
+  https://www.unicode.org/Public/12.1.0/ucd/extracted/DerivedCombiningClass.txt
+  https://www.unicode.org/Public/12.1.0/ucd/DerivedCoreProperties.txt
+  https://www.unicode.org/Public/12.1.0/ucd/NormalizationCorrections.txt
+  https://www.unicode.org/Public/12.1.0/ucd/NormalizationTest.txt
+  https://www.unicode.org/Public/12.1.0/ucd/UnicodeData.txt
+
+md5sums (verify by running "md5sum -c README.utf8data"):
+
+  900e76da1d822a160fd6b8c0b1d70094  CaseFolding.txt
+  131256380bff4fea8ad4a851616f2f10  DerivedAge.txt
+  e731a4089b30002144e107e3d6f8d1fa  DerivedCombiningClass.txt
+  a47c9fbd7ff92a9b261ba9831e68778a  DerivedCoreProperties.txt
+  fcab6dad15e440879d92f315978f93d3  NormalizationCorrections.txt
+  f9ff1c55a60decf436100f791b44aa98  NormalizationTest.txt
+  755f6af699f8c8d2d958da411f78f6c6  UnicodeData.txt
+
+sha1sums (verify by running "sha1sum -c README.utf8data"):
+
+  dc9245f6803c4ac99555c361f5052e0b13eb779b  CaseFolding.txt
+  3281104f237184cdb5d869e86eb8573678ada7da  DerivedAge.txt
+  2f5f995ccb96e0fa84b15151b35d5e2681535175  DerivedCombiningClass.txt
+  5b8698a3fcd5018e1987f296b02e2c17e696415e  DerivedCoreProperties.txt
+  cd83935fbc012345d8792d2c704f69497e753835  NormalizationCorrections.txt
+  ea419aae505b337b0d99a83fa83fe58ddff7c19f  NormalizationTest.txt
+  dc973c0fc93d6f09d9ab9f70d1c9f89c447f0526  UnicodeData.txt
+
+
+To update to the newer version of the Unicode standard, the latest
+released version of the UCD can be found here:
+
+  http://www.unicode.org/Public/UCD/latest/
+
+Then, build under fs/unicode/ with REGENERATE_UTF8DATA=1:
+
+	make REGENERATE_UTF8DATA=1 fs/unicode/
+
+After sanity checking the newly generated utf8data.c file (the
+version generated from the 12.1.0 UCD should be 4,109 lines long, and
+have a total size of 324k) and/or comparing it with the older version
+of utf8data.c_shipped, rename it to utf8data.c_shipped.
+
+If you are a kernel developer updating to a newer version of the
+Unicode Character Database, please update this README.utf8data file
+with the version of the UCD that was used, the md5sum and sha1sums of
+the *.txt files, before checking in the new versions of the utf8data.c
+and README.utf8data files.

linux-7.2.3/fs/unicode/tests/.kunitconfig

@@ -0,0 +1,3 @@
+CONFIG_KUNIT=y
+CONFIG_UNICODE=y
+CONFIG_UNICODE_NORMALIZATION_KUNIT_TEST=y

linux-7.2.3/fs/unicode/tests/utf8_kunit.c

@@ -0,0 +1,297 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * KUnit tests for utf-8 support.
+ *
+ * Copyright 2017 Collabora Ltd.
+ */
+
+#include <linux/unicode.h>
+#include <kunit/test.h>
+
+#include "../utf8n.h"
+
+static const struct {
+	/* UTF-8 strings in this vector _must_ be NULL-terminated. */
+	unsigned char str[10];
+	unsigned char dec[10];
+} nfdi_test_data[] = {
+	/* Trivial sequence */
+	{
+		/* "ABba" decomposes to itself */
+		.str = "aBba",
+		.dec = "aBba",
+	},
+	/* Simple equivalent sequences */
+	{
+               /* 'VULGAR FRACTION ONE QUARTER' cannot decompose to
+                  'NUMBER 1' + 'FRACTION SLASH' + 'NUMBER 4' on
+                  canonical decomposition */
+               .str = {0xc2, 0xbc, 0x00},
+	       .dec = {0xc2, 0xbc, 0x00},
+	},
+	{
+		/* 'LATIN SMALL LETTER A WITH DIAERESIS' decomposes to
+		   'LETTER A' + 'COMBINING DIAERESIS' */
+		.str = {0xc3, 0xa4, 0x00},
+		.dec = {0x61, 0xcc, 0x88, 0x00},
+	},
+	{
+		/* 'LATIN SMALL LETTER LJ' can't decompose to
+		   'LETTER L' + 'LETTER J' on canonical decomposition */
+		.str = {0xC7, 0x89, 0x00},
+		.dec = {0xC7, 0x89, 0x00},
+	},
+	{
+		/* GREEK ANO TELEIA decomposes to MIDDLE DOT */
+		.str = {0xCE, 0x87, 0x00},
+		.dec = {0xC2, 0xB7, 0x00}
+	},
+	/* Canonical ordering */
+	{
+		/* A + 'COMBINING ACUTE ACCENT' + 'COMBINING OGONEK' decomposes
+		   to A + 'COMBINING OGONEK' + 'COMBINING ACUTE ACCENT' */
+		.str = {0x41, 0xcc, 0x81, 0xcc, 0xa8, 0x0},
+		.dec = {0x41, 0xcc, 0xa8, 0xcc, 0x81, 0x0},
+	},
+	{
+		/* 'LATIN SMALL LETTER A WITH DIAERESIS' + 'COMBINING OGONEK'
+		   decomposes to
+		   'LETTER A' + 'COMBINING OGONEK' + 'COMBINING DIAERESIS' */
+		.str = {0xc3, 0xa4, 0xCC, 0xA8, 0x00},
+
+		.dec = {0x61, 0xCC, 0xA8, 0xcc, 0x88, 0x00},
+	},
+
+};
+
+static const struct {
+	/* UTF-8 strings in this vector _must_ be NULL-terminated. */
+	unsigned char str[30];
+	unsigned char ncf[30];
+} nfdicf_test_data[] = {
+	/* Trivial sequences */
+	{
+		/* "ABba" folds to lowercase */
+		.str = {0x41, 0x42, 0x62, 0x61, 0x00},
+		.ncf = {0x61, 0x62, 0x62, 0x61, 0x00},
+	},
+	{
+		/* All ASCII folds to lower-case */
+		.str = "ABCDEFGHIJKLMNOPQRSTUVWXYZ0.1",
+		.ncf = "abcdefghijklmnopqrstuvwxyz0.1",
+	},
+	{
+		/* LATIN SMALL LETTER SHARP S folds to
+		   LATIN SMALL LETTER S + LATIN SMALL LETTER S */
+		.str = {0xc3, 0x9f, 0x00},
+		.ncf = {0x73, 0x73, 0x00},
+	},
+	{
+		/* LATIN CAPITAL LETTER A WITH RING ABOVE folds to
+		   LATIN SMALL LETTER A + COMBINING RING ABOVE */
+		.str = {0xC3, 0x85, 0x00},
+		.ncf = {0x61, 0xcc, 0x8a, 0x00},
+	},
+	/* Introduced by UTF-8.0.0. */
+	/* Cherokee letters are interesting test-cases because they fold
+	   to upper-case.  Before 8.0.0, Cherokee lowercase were
+	   undefined, thus, the folding from LC is not stable between
+	   7.0.0 -> 8.0.0, but it is from UC. */
+	{
+		/* CHEROKEE SMALL LETTER A folds to CHEROKEE LETTER A */
+		.str = {0xea, 0xad, 0xb0, 0x00},
+		.ncf = {0xe1, 0x8e, 0xa0, 0x00},
+	},
+	{
+		/* CHEROKEE SMALL LETTER YE folds to CHEROKEE LETTER YE */
+		.str = {0xe1, 0x8f, 0xb8, 0x00},
+		.ncf = {0xe1, 0x8f, 0xb0, 0x00},
+	},
+	{
+		/* OLD HUNGARIAN CAPITAL LETTER AMB folds to
+		   OLD HUNGARIAN SMALL LETTER AMB */
+		.str = {0xf0, 0x90, 0xb2, 0x83, 0x00},
+		.ncf = {0xf0, 0x90, 0xb3, 0x83, 0x00},
+	},
+	/* Introduced by UTF-9.0.0. */
+	{
+		/* OSAGE CAPITAL LETTER CHA folds to
+		   OSAGE SMALL LETTER CHA */
+		.str = {0xf0, 0x90, 0x92, 0xb5, 0x00},
+		.ncf = {0xf0, 0x90, 0x93, 0x9d, 0x00},
+	},
+	{
+		/* LATIN CAPITAL LETTER SMALL CAPITAL I folds to
+		   LATIN LETTER SMALL CAPITAL I */
+		.str = {0xea, 0x9e, 0xae, 0x00},
+		.ncf = {0xc9, 0xaa, 0x00},
+	},
+	/* Introduced by UTF-11.0.0. */
+	{
+		/* GEORGIAN SMALL LETTER AN folds to GEORGIAN MTAVRULI
+		   CAPITAL LETTER AN */
+		.str = {0xe1, 0xb2, 0x90, 0x00},
+		.ncf = {0xe1, 0x83, 0x90, 0x00},
+	}
+};
+
+static ssize_t utf8len(const struct unicode_map *um, enum utf8_normalization n,
+		const char *s)
+{
+	return utf8nlen(um, n, s, (size_t)-1);
+}
+
+static int utf8cursor(struct utf8cursor *u8c, const struct unicode_map *um,
+		enum utf8_normalization n, const char *s)
+{
+	return utf8ncursor(u8c, um, n, s, (unsigned int)-1);
+}
+
+static void check_utf8_nfdi(struct kunit *test)
+{
+	int i;
+	struct utf8cursor u8c;
+	struct unicode_map *um = test->priv;
+
+	for (i = 0; i < ARRAY_SIZE(nfdi_test_data); i++) {
+		int len = strlen(nfdi_test_data[i].str);
+		int nlen = strlen(nfdi_test_data[i].dec);
+		int j = 0;
+		unsigned char c;
+		int ret;
+
+		KUNIT_EXPECT_EQ(test, utf8len(um, UTF8_NFDI, nfdi_test_data[i].str), nlen);
+		KUNIT_EXPECT_EQ(test, utf8nlen(um, UTF8_NFDI, nfdi_test_data[i].str, len),
+				nlen);
+
+
+		ret = utf8cursor(&u8c, um, UTF8_NFDI, nfdi_test_data[i].str);
+		KUNIT_EXPECT_TRUE_MSG(test, ret >= 0, "Can't create cursor\n");
+
+		while ((c = utf8byte(&u8c)) > 0) {
+			KUNIT_EXPECT_EQ_MSG(test, c, nfdi_test_data[i].dec[j],
+					    "Unexpected byte 0x%x should be 0x%x\n",
+					    c, nfdi_test_data[i].dec[j]);
+			j++;
+		}
+
+		KUNIT_EXPECT_EQ(test, j, nlen);
+	}
+}
+
+static void check_utf8_nfdicf(struct kunit *test)
+{
+	int i;
+	struct utf8cursor u8c;
+	struct unicode_map *um = test->priv;
+
+	for (i = 0; i < ARRAY_SIZE(nfdicf_test_data); i++) {
+		int len = strlen(nfdicf_test_data[i].str);
+		int nlen = strlen(nfdicf_test_data[i].ncf);
+		int j = 0;
+		int ret;
+		unsigned char c;
+
+		KUNIT_EXPECT_EQ(test, utf8len(um, UTF8_NFDICF, nfdicf_test_data[i].str),
+				nlen);
+		KUNIT_EXPECT_EQ(test, utf8nlen(um, UTF8_NFDICF, nfdicf_test_data[i].str, len),
+				nlen);
+
+		ret = utf8cursor(&u8c, um, UTF8_NFDICF, nfdicf_test_data[i].str);
+		KUNIT_EXPECT_TRUE_MSG(test, ret >= 0, "Can't create cursor\n");
+
+		while ((c = utf8byte(&u8c)) > 0) {
+			KUNIT_EXPECT_EQ_MSG(test, c, nfdicf_test_data[i].ncf[j],
+					    "Unexpected byte 0x%x should be 0x%x\n",
+					    c, nfdicf_test_data[i].ncf[j]);
+			j++;
+		}
+
+		KUNIT_EXPECT_EQ(test, j, nlen);
+	}
+}
+
+static void check_utf8_comparisons(struct kunit *test)
+{
+	int i;
+	struct unicode_map *um = test->priv;
+
+	for (i = 0; i < ARRAY_SIZE(nfdi_test_data); i++) {
+		const struct qstr s1 = {.name = nfdi_test_data[i].str,
+					.len = sizeof(nfdi_test_data[i].str)};
+		const struct qstr s2 = {.name = nfdi_test_data[i].dec,
+					.len = sizeof(nfdi_test_data[i].dec)};
+
+		/* strncmp returns 0 when strings are equal */
+		KUNIT_EXPECT_TRUE_MSG(test, utf8_strncmp(um, &s1, &s2) == 0,
+				    "%s %s comparison mismatch\n", s1.name, s2.name);
+	}
+
+	for (i = 0; i < ARRAY_SIZE(nfdicf_test_data); i++) {
+		const struct qstr s1 = {.name = nfdicf_test_data[i].str,
+					.len = sizeof(nfdicf_test_data[i].str)};
+		const struct qstr s2 = {.name = nfdicf_test_data[i].ncf,
+					.len = sizeof(nfdicf_test_data[i].ncf)};
+
+		/* strncasecmp returns 0 when strings are equal */
+		KUNIT_EXPECT_TRUE_MSG(test, utf8_strncasecmp(um, &s1, &s2) == 0,
+				    "%s %s comparison mismatch\n", s1.name, s2.name);
+	}
+}
+
+static void check_supported_versions(struct kunit *test)
+{
+	struct unicode_map *um = test->priv;
+	/* Unicode 7.0.0 should be supported. */
+	KUNIT_EXPECT_TRUE(test, utf8version_is_supported(um, UNICODE_AGE(7, 0, 0)));
+
+	/* Unicode 9.0.0 should be supported. */
+	KUNIT_EXPECT_TRUE(test, utf8version_is_supported(um, UNICODE_AGE(9, 0, 0)));
+
+	/* Unicode 1x.0.0 (the latest version) should be supported. */
+	KUNIT_EXPECT_TRUE(test, utf8version_is_supported(um, UTF8_LATEST));
+
+	/* Next versions don't exist. */
+	KUNIT_EXPECT_FALSE(test, utf8version_is_supported(um, UNICODE_AGE(13, 0, 0)));
+	KUNIT_EXPECT_FALSE(test, utf8version_is_supported(um, UNICODE_AGE(0, 0, 0)));
+	KUNIT_EXPECT_FALSE(test, utf8version_is_supported(um, UNICODE_AGE(-1, -1, -1)));
+}
+
+static struct kunit_case unicode_normalization_test_cases[] = {
+	KUNIT_CASE(check_supported_versions),
+	KUNIT_CASE(check_utf8_comparisons),
+	KUNIT_CASE(check_utf8_nfdicf),
+	KUNIT_CASE(check_utf8_nfdi),
+	{}
+};
+
+static int init_test_ucd(struct kunit *test)
+{
+	struct unicode_map *um = utf8_load(UTF8_LATEST);
+
+	test->priv = um;
+
+	KUNIT_EXPECT_EQ_MSG(test, IS_ERR(um), 0,
+			    "%s: Unable to load utf8 table.\n", __func__);
+
+	return 0;
+}
+
+static void exit_test_ucd(struct kunit *test)
+{
+	utf8_unload(test->priv);
+}
+
+static struct kunit_suite unicode_normalization_test_suite = {
+	.name = "unicode_normalization",
+	.test_cases = unicode_normalization_test_cases,
+	.init = init_test_ucd,
+	.exit = exit_test_ucd,
+};
+
+kunit_test_suite(unicode_normalization_test_suite);
+
+
+MODULE_AUTHOR("Gabriel Krisman Bertazi <krisman@collabora.co.uk>");
+MODULE_DESCRIPTION("KUnit tests for utf-8 support.");
+MODULE_LICENSE("GPL");

linux-7.2.3/fs/unicode/utf8-core.c

@@ -0,0 +1,242 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/slab.h>
+#include <linux/parser.h>
+#include <linux/errno.h>
+#include <linux/stringhash.h>
+
+#include "utf8n.h"
+
+int utf8_validate(const struct unicode_map *um, const struct qstr *str)
+{
+	if (utf8nlen(um, UTF8_NFDI, str->name, str->len) < 0)
+		return -1;
+	return 0;
+}
+EXPORT_SYMBOL(utf8_validate);
+
+int utf8_strncmp(const struct unicode_map *um,
+		 const struct qstr *s1, const struct qstr *s2)
+{
+	struct utf8cursor cur1, cur2;
+	int c1, c2;
+
+	if (utf8ncursor(&cur1, um, UTF8_NFDI, s1->name, s1->len) < 0)
+		return -EINVAL;
+
+	if (utf8ncursor(&cur2, um, UTF8_NFDI, s2->name, s2->len) < 0)
+		return -EINVAL;
+
+	do {
+		c1 = utf8byte(&cur1);
+		c2 = utf8byte(&cur2);
+
+		if (c1 < 0 || c2 < 0)
+			return -EINVAL;
+		if (c1 != c2)
+			return 1;
+	} while (c1);
+
+	return 0;
+}
+EXPORT_SYMBOL(utf8_strncmp);
+
+int utf8_strncasecmp(const struct unicode_map *um,
+		     const struct qstr *s1, const struct qstr *s2)
+{
+	struct utf8cursor cur1, cur2;
+	int c1, c2;
+
+	if (utf8ncursor(&cur1, um, UTF8_NFDICF, s1->name, s1->len) < 0)
+		return -EINVAL;
+
+	if (utf8ncursor(&cur2, um, UTF8_NFDICF, s2->name, s2->len) < 0)
+		return -EINVAL;
+
+	do {
+		c1 = utf8byte(&cur1);
+		c2 = utf8byte(&cur2);
+
+		if (c1 < 0 || c2 < 0)
+			return -EINVAL;
+		if (c1 != c2)
+			return 1;
+	} while (c1);
+
+	return 0;
+}
+EXPORT_SYMBOL(utf8_strncasecmp);
+
+/* String cf is expected to be a valid UTF-8 casefolded
+ * string.
+ */
+int utf8_strncasecmp_folded(const struct unicode_map *um,
+			    const struct qstr *cf,
+			    const struct qstr *s1)
+{
+	struct utf8cursor cur1;
+	int c1, c2;
+	int i = 0;
+
+	if (utf8ncursor(&cur1, um, UTF8_NFDICF, s1->name, s1->len) < 0)
+		return -EINVAL;
+
+	do {
+		c1 = utf8byte(&cur1);
+		c2 = cf->name[i++];
+		if (c1 < 0)
+			return -EINVAL;
+		if (c1 != c2)
+			return 1;
+	} while (c1);
+
+	return 0;
+}
+EXPORT_SYMBOL(utf8_strncasecmp_folded);
+
+int utf8_casefold(const struct unicode_map *um, const struct qstr *str,
+		  unsigned char *dest, size_t dlen)
+{
+	struct utf8cursor cur;
+	size_t nlen = 0;
+
+	if (utf8ncursor(&cur, um, UTF8_NFDICF, str->name, str->len) < 0)
+		return -EINVAL;
+
+	for (nlen = 0; nlen < dlen; nlen++) {
+		int c = utf8byte(&cur);
+
+		dest[nlen] = c;
+		if (!c)
+			return nlen;
+		if (c == -1)
+			break;
+	}
+	return -EINVAL;
+}
+EXPORT_SYMBOL(utf8_casefold);
+
+int utf8_casefold_hash(const struct unicode_map *um, const void *salt,
+		       struct qstr *str)
+{
+	struct utf8cursor cur;
+	int c;
+	unsigned long hash = init_name_hash(salt);
+
+	if (utf8ncursor(&cur, um, UTF8_NFDICF, str->name, str->len) < 0)
+		return -EINVAL;
+
+	while ((c = utf8byte(&cur))) {
+		if (c < 0)
+			return -EINVAL;
+		hash = partial_name_hash((unsigned char)c, hash);
+	}
+	str->hash = end_name_hash(hash);
+	return 0;
+}
+EXPORT_SYMBOL(utf8_casefold_hash);
+
+int utf8_normalize(const struct unicode_map *um, const struct qstr *str,
+		   unsigned char *dest, size_t dlen)
+{
+	struct utf8cursor cur;
+	ssize_t nlen = 0;
+
+	if (utf8ncursor(&cur, um, UTF8_NFDI, str->name, str->len) < 0)
+		return -EINVAL;
+
+	for (nlen = 0; nlen < dlen; nlen++) {
+		int c = utf8byte(&cur);
+
+		dest[nlen] = c;
+		if (!c)
+			return nlen;
+		if (c == -1)
+			break;
+	}
+	return -EINVAL;
+}
+EXPORT_SYMBOL(utf8_normalize);
+
+static const struct utf8data *find_table_version(const struct utf8data *table,
+		size_t nr_entries, unsigned int version)
+{
+	size_t i = nr_entries - 1;
+
+	while (version < table[i].maxage)
+		i--;
+	if (version > table[i].maxage)
+		return NULL;
+	return &table[i];
+}
+
+struct unicode_map *utf8_load(unsigned int version)
+{
+	struct unicode_map *um;
+
+	um = kzalloc_obj(struct unicode_map);
+	if (!um)
+		return ERR_PTR(-ENOMEM);
+	um->version = version;
+
+	um->tables = symbol_request(utf8_data_table);
+	if (!um->tables)
+		goto out_free_um;
+
+	if (!utf8version_is_supported(um, version))
+		goto out_symbol_put;
+	um->ntab[UTF8_NFDI] = find_table_version(um->tables->utf8nfdidata,
+			um->tables->utf8nfdidata_size, um->version);
+	if (!um->ntab[UTF8_NFDI])
+		goto out_symbol_put;
+	um->ntab[UTF8_NFDICF] = find_table_version(um->tables->utf8nfdicfdata,
+			um->tables->utf8nfdicfdata_size, um->version);
+	if (!um->ntab[UTF8_NFDICF])
+		goto out_symbol_put;
+	return um;
+
+out_symbol_put:
+	symbol_put(utf8_data_table);
+out_free_um:
+	kfree(um);
+	return ERR_PTR(-EINVAL);
+}
+EXPORT_SYMBOL(utf8_load);
+
+void utf8_unload(struct unicode_map *um)
+{
+	if (um) {
+		symbol_put(utf8_data_table);
+		kfree(um);
+	}
+}
+EXPORT_SYMBOL(utf8_unload);
+
+/**
+ * utf8_parse_version - Parse a UTF-8 version number from a string
+ *
+ * @version: input string
+ *
+ * Returns the parsed version on success, negative code on error
+ */
+int utf8_parse_version(char *version)
+{
+	substring_t args[3];
+	unsigned int maj, min, rev;
+	static const struct match_token token[] = {
+		{1, "%d.%d.%d"},
+		{0, NULL}
+	};
+
+	if (match_token(version, token, args) != 1)
+		return -EINVAL;
+
+	if (match_int(&args[0], &maj) || match_int(&args[1], &min) ||
+	    match_int(&args[2], &rev))
+		return -EINVAL;
+
+	return UNICODE_AGE(maj, min, rev);
+}
+EXPORT_SYMBOL(utf8_parse_version);

linux-7.2.3/fs/unicode/utf8-norm.c

@@ -0,0 +1,594 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2014 SGI.
+ * All rights reserved.
+ */
+
+#include "utf8n.h"
+
+int utf8version_is_supported(const struct unicode_map *um, unsigned int version)
+{
+	int i = um->tables->utf8agetab_size - 1;
+
+	while (i >= 0 && um->tables->utf8agetab[i] != 0) {
+		if (version == um->tables->utf8agetab[i])
+			return 1;
+		i--;
+	}
+	return 0;
+}
+
+/*
+ * UTF-8 valid ranges.
+ *
+ * The UTF-8 encoding spreads the bits of a 32bit word over several
+ * bytes. This table gives the ranges that can be held and how they'd
+ * be represented.
+ *
+ * 0x00000000 0x0000007F: 0xxxxxxx
+ * 0x00000000 0x000007FF: 110xxxxx 10xxxxxx
+ * 0x00000000 0x0000FFFF: 1110xxxx 10xxxxxx 10xxxxxx
+ * 0x00000000 0x001FFFFF: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
+ * 0x00000000 0x03FFFFFF: 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
+ * 0x00000000 0x7FFFFFFF: 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
+ *
+ * There is an additional requirement on UTF-8, in that only the
+ * shortest representation of a 32bit value is to be used.  A decoder
+ * must not decode sequences that do not satisfy this requirement.
+ * Thus the allowed ranges have a lower bound.
+ *
+ * 0x00000000 0x0000007F: 0xxxxxxx
+ * 0x00000080 0x000007FF: 110xxxxx 10xxxxxx
+ * 0x00000800 0x0000FFFF: 1110xxxx 10xxxxxx 10xxxxxx
+ * 0x00010000 0x001FFFFF: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
+ * 0x00200000 0x03FFFFFF: 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
+ * 0x04000000 0x7FFFFFFF: 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
+ *
+ * Actual unicode characters are limited to the range 0x0 - 0x10FFFF,
+ * 17 planes of 65536 values.  This limits the sequences actually seen
+ * even more, to just the following.
+ *
+ *          0 -     0x7F: 0                   - 0x7F
+ *       0x80 -    0x7FF: 0xC2 0x80           - 0xDF 0xBF
+ *      0x800 -   0xFFFF: 0xE0 0xA0 0x80      - 0xEF 0xBF 0xBF
+ *    0x10000 - 0x10FFFF: 0xF0 0x90 0x80 0x80 - 0xF4 0x8F 0xBF 0xBF
+ *
+ * Within those ranges the surrogates 0xD800 - 0xDFFF are not allowed.
+ *
+ * Note that the longest sequence seen with valid usage is 4 bytes,
+ * the same a single UTF-32 character.  This makes the UTF-8
+ * representation of Unicode strictly smaller than UTF-32.
+ *
+ * The shortest sequence requirement was introduced by:
+ *    Corrigendum #1: UTF-8 Shortest Form
+ * It can be found here:
+ *    http://www.unicode.org/versions/corrigendum1.html
+ *
+ */
+
+/*
+ * Return the number of bytes used by the current UTF-8 sequence.
+ * Assumes the input points to the first byte of a valid UTF-8
+ * sequence.
+ */
+static inline int utf8clen(const char *s)
+{
+	unsigned char c = *s;
+
+	return 1 + (c >= 0xC0) + (c >= 0xE0) + (c >= 0xF0);
+}
+
+/*
+ * Decode a 3-byte UTF-8 sequence.
+ */
+static unsigned int
+utf8decode3(const char *str)
+{
+	unsigned int		uc;
+
+	uc = *str++ & 0x0F;
+	uc <<= 6;
+	uc |= *str++ & 0x3F;
+	uc <<= 6;
+	uc |= *str++ & 0x3F;
+
+	return uc;
+}
+
+/*
+ * Encode a 3-byte UTF-8 sequence.
+ */
+static int
+utf8encode3(char *str, unsigned int val)
+{
+	str[2] = (val & 0x3F) | 0x80;
+	val >>= 6;
+	str[1] = (val & 0x3F) | 0x80;
+	val >>= 6;
+	str[0] = val | 0xE0;
+
+	return 3;
+}
+
+/*
+ * utf8trie_t
+ *
+ * A compact binary tree, used to decode UTF-8 characters.
+ *
+ * Internal nodes are one byte for the node itself, and up to three
+ * bytes for an offset into the tree.  The first byte contains the
+ * following information:
+ *  NEXTBYTE  - flag        - advance to next byte if set
+ *  BITNUM    - 3 bit field - the bit number to tested
+ *  OFFLEN    - 2 bit field - number of bytes in the offset
+ * if offlen == 0 (non-branching node)
+ *  RIGHTPATH - 1 bit field - set if the following node is for the
+ *                            right-hand path (tested bit is set)
+ *  TRIENODE  - 1 bit field - set if the following node is an internal
+ *                            node, otherwise it is a leaf node
+ * if offlen != 0 (branching node)
+ *  LEFTNODE  - 1 bit field - set if the left-hand node is internal
+ *  RIGHTNODE - 1 bit field - set if the right-hand node is internal
+ *
+ * Due to the way utf8 works, there cannot be branching nodes with
+ * NEXTBYTE set, and moreover those nodes always have a righthand
+ * descendant.
+ */
+typedef const unsigned char utf8trie_t;
+#define BITNUM		0x07
+#define NEXTBYTE	0x08
+#define OFFLEN		0x30
+#define OFFLEN_SHIFT	4
+#define RIGHTPATH	0x40
+#define TRIENODE	0x80
+#define RIGHTNODE	0x40
+#define LEFTNODE	0x80
+
+/*
+ * utf8leaf_t
+ *
+ * The leaves of the trie are embedded in the trie, and so the same
+ * underlying datatype: unsigned char.
+ *
+ * leaf[0]: The unicode version, stored as a generation number that is
+ *          an index into ->utf8agetab[].  With this we can filter code
+ *          points based on the unicode version in which they were
+ *          defined.  The CCC of a non-defined code point is 0.
+ * leaf[1]: Canonical Combining Class. During normalization, we need
+ *          to do a stable sort into ascending order of all characters
+ *          with a non-zero CCC that occur between two characters with
+ *          a CCC of 0, or at the begin or end of a string.
+ *          The unicode standard guarantees that all CCC values are
+ *          between 0 and 254 inclusive, which leaves 255 available as
+ *          a special value.
+ *          Code points with CCC 0 are known as stoppers.
+ * leaf[2]: Decomposition. If leaf[1] == 255, then leaf[2] is the
+ *          start of a NUL-terminated string that is the decomposition
+ *          of the character.
+ *          The CCC of a decomposable character is the same as the CCC
+ *          of the first character of its decomposition.
+ *          Some characters decompose as the empty string: these are
+ *          characters with the Default_Ignorable_Code_Point property.
+ *          These do affect normalization, as they all have CCC 0.
+ *
+ * The decompositions in the trie have been fully expanded, with the
+ * exception of Hangul syllables, which are decomposed algorithmically.
+ *
+ * Casefolding, if applicable, is also done using decompositions.
+ *
+ * The trie is constructed in such a way that leaves exist for all
+ * UTF-8 sequences that match the criteria from the "UTF-8 valid
+ * ranges" comment above, and only for those sequences.  Therefore a
+ * lookup in the trie can be used to validate the UTF-8 input.
+ */
+typedef const unsigned char utf8leaf_t;
+
+#define LEAF_GEN(LEAF)	((LEAF)[0])
+#define LEAF_CCC(LEAF)	((LEAF)[1])
+#define LEAF_STR(LEAF)	((const char *)((LEAF) + 2))
+
+#define MINCCC		(0)
+#define MAXCCC		(254)
+#define STOPPER		(0)
+#define	DECOMPOSE	(255)
+
+/* Marker for hangul syllable decomposition. */
+#define HANGUL		((char)(255))
+/* Size of the synthesized leaf used for Hangul syllable decomposition. */
+#define UTF8HANGULLEAF	(12)
+
+/*
+ * Hangul decomposition (algorithm from Section 3.12 of Unicode 6.3.0)
+ *
+ * AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;;
+ * D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;;
+ *
+ * SBase = 0xAC00
+ * LBase = 0x1100
+ * VBase = 0x1161
+ * TBase = 0x11A7
+ * LCount = 19
+ * VCount = 21
+ * TCount = 28
+ * NCount = 588 (VCount * TCount)
+ * SCount = 11172 (LCount * NCount)
+ *
+ * Decomposition:
+ *   SIndex = s - SBase
+ *
+ * LV (Canonical/Full)
+ *   LIndex = SIndex / NCount
+ *   VIndex = (Sindex % NCount) / TCount
+ *   LPart = LBase + LIndex
+ *   VPart = VBase + VIndex
+ *
+ * LVT (Canonical)
+ *   LVIndex = (SIndex / TCount) * TCount
+ *   TIndex = (Sindex % TCount)
+ *   LVPart = SBase + LVIndex
+ *   TPart = TBase + TIndex
+ *
+ * LVT (Full)
+ *   LIndex = SIndex / NCount
+ *   VIndex = (Sindex % NCount) / TCount
+ *   TIndex = (Sindex % TCount)
+ *   LPart = LBase + LIndex
+ *   VPart = VBase + VIndex
+ *   if (TIndex == 0) {
+ *          d = <LPart, VPart>
+ *   } else {
+ *          TPart = TBase + TIndex
+ *          d = <LPart, TPart, VPart>
+ *   }
+ */
+
+/* Constants */
+#define SB	(0xAC00)
+#define LB	(0x1100)
+#define VB	(0x1161)
+#define TB	(0x11A7)
+#define LC	(19)
+#define VC	(21)
+#define TC	(28)
+#define NC	(VC * TC)
+#define SC	(LC * NC)
+
+/* Algorithmic decomposition of hangul syllable. */
+static utf8leaf_t *
+utf8hangul(const char *str, unsigned char *hangul)
+{
+	unsigned int	si;
+	unsigned int	li;
+	unsigned int	vi;
+	unsigned int	ti;
+	unsigned char	*h;
+
+	/* Calculate the SI, LI, VI, and TI values. */
+	si = utf8decode3(str) - SB;
+	li = si / NC;
+	vi = (si % NC) / TC;
+	ti = si % TC;
+
+	/* Fill in base of leaf. */
+	h = hangul;
+	LEAF_GEN(h) = 2;
+	LEAF_CCC(h) = DECOMPOSE;
+	h += 2;
+
+	/* Add LPart, a 3-byte UTF-8 sequence. */
+	h += utf8encode3((char *)h, li + LB);
+
+	/* Add VPart, a 3-byte UTF-8 sequence. */
+	h += utf8encode3((char *)h, vi + VB);
+
+	/* Add TPart if required, also a 3-byte UTF-8 sequence. */
+	if (ti)
+		h += utf8encode3((char *)h, ti + TB);
+
+	/* Terminate string. */
+	h[0] = '\0';
+
+	return hangul;
+}
+
+/*
+ * Use trie to scan s, touching at most len bytes.
+ * Returns the leaf if one exists, NULL otherwise.
+ *
+ * A non-NULL return guarantees that the UTF-8 sequence starting at s
+ * is well-formed and corresponds to a known unicode code point.  The
+ * shorthand for this will be "is valid UTF-8 unicode".
+ */
+static utf8leaf_t *utf8nlookup(const struct unicode_map *um,
+		enum utf8_normalization n, unsigned char *hangul, const char *s,
+		size_t len)
+{
+	utf8trie_t	*trie = um->tables->utf8data + um->ntab[n]->offset;
+	int		offlen;
+	int		offset;
+	int		mask;
+	int		node;
+
+	if (len == 0)
+		return NULL;
+
+	node = 1;
+	while (node) {
+		offlen = (*trie & OFFLEN) >> OFFLEN_SHIFT;
+		if (*trie & NEXTBYTE) {
+			if (--len == 0)
+				return NULL;
+			s++;
+		}
+		mask = 1 << (*trie & BITNUM);
+		if (*s & mask) {
+			/* Right leg */
+			if (offlen) {
+				/* Right node at offset of trie */
+				node = (*trie & RIGHTNODE);
+				offset = trie[offlen];
+				while (--offlen) {
+					offset <<= 8;
+					offset |= trie[offlen];
+				}
+				trie += offset;
+			} else if (*trie & RIGHTPATH) {
+				/* Right node after this node */
+				node = (*trie & TRIENODE);
+				trie++;
+			} else {
+				/* No right node. */
+				return NULL;
+			}
+		} else {
+			/* Left leg */
+			if (offlen) {
+				/* Left node after this node. */
+				node = (*trie & LEFTNODE);
+				trie += offlen + 1;
+			} else if (*trie & RIGHTPATH) {
+				/* No left node. */
+				return NULL;
+			} else {
+				/* Left node after this node */
+				node = (*trie & TRIENODE);
+				trie++;
+			}
+		}
+	}
+	/*
+	 * Hangul decomposition is done algorithmically. These are the
+	 * codepoints >= 0xAC00 and <= 0xD7A3. Their UTF-8 encoding is
+	 * always 3 bytes long, so s has been advanced twice, and the
+	 * start of the sequence is at s-2.
+	 */
+	if (LEAF_CCC(trie) == DECOMPOSE && LEAF_STR(trie)[0] == HANGUL)
+		trie = utf8hangul(s - 2, hangul);
+	return trie;
+}
+
+/*
+ * Use trie to scan s.
+ * Returns the leaf if one exists, NULL otherwise.
+ *
+ * Forwards to utf8nlookup().
+ */
+static utf8leaf_t *utf8lookup(const struct unicode_map *um,
+		enum utf8_normalization n, unsigned char *hangul, const char *s)
+{
+	return utf8nlookup(um, n, hangul, s, (size_t)-1);
+}
+
+/*
+ * Length of the normalization of s, touch at most len bytes.
+ * Return -1 if s is not valid UTF-8 unicode.
+ */
+ssize_t utf8nlen(const struct unicode_map *um, enum utf8_normalization n,
+		const char *s, size_t len)
+{
+	utf8leaf_t	*leaf;
+	size_t		ret = 0;
+	unsigned char	hangul[UTF8HANGULLEAF];
+
+	while (len && *s) {
+		leaf = utf8nlookup(um, n, hangul, s, len);
+		if (!leaf)
+			return -1;
+		if (um->tables->utf8agetab[LEAF_GEN(leaf)] >
+		    um->ntab[n]->maxage)
+			ret += utf8clen(s);
+		else if (LEAF_CCC(leaf) == DECOMPOSE)
+			ret += strlen(LEAF_STR(leaf));
+		else
+			ret += utf8clen(s);
+		len -= utf8clen(s);
+		s += utf8clen(s);
+	}
+	return ret;
+}
+
+/*
+ * Set up an utf8cursor for use by utf8byte().
+ *
+ *   u8c    : pointer to cursor.
+ *   data   : const struct utf8data to use for normalization.
+ *   s      : string.
+ *   len    : length of s.
+ *
+ * Returns -1 on error, 0 on success.
+ */
+int utf8ncursor(struct utf8cursor *u8c, const struct unicode_map *um,
+		enum utf8_normalization n, const char *s, size_t len)
+{
+	if (!s)
+		return -1;
+	u8c->um = um;
+	u8c->n = n;
+	u8c->s = s;
+	u8c->p = NULL;
+	u8c->ss = NULL;
+	u8c->sp = NULL;
+	u8c->len = len;
+	u8c->slen = 0;
+	u8c->ccc = STOPPER;
+	u8c->nccc = STOPPER;
+	/* Check we didn't clobber the maximum length. */
+	if (u8c->len != len)
+		return -1;
+	/* The first byte of s may not be an utf8 continuation. */
+	if (len > 0 && (*s & 0xC0) == 0x80)
+		return -1;
+	return 0;
+}
+
+/*
+ * Get one byte from the normalized form of the string described by u8c.
+ *
+ * Returns the byte cast to an unsigned char on succes, and -1 on failure.
+ *
+ * The cursor keeps track of the location in the string in u8c->s.
+ * When a character is decomposed, the current location is stored in
+ * u8c->p, and u8c->s is set to the start of the decomposition. Note
+ * that bytes from a decomposition do not count against u8c->len.
+ *
+ * Characters are emitted if they match the current CCC in u8c->ccc.
+ * Hitting end-of-string while u8c->ccc == STOPPER means we're done,
+ * and the function returns 0 in that case.
+ *
+ * Sorting by CCC is done by repeatedly scanning the string.  The
+ * values of u8c->s and u8c->p are stored in u8c->ss and u8c->sp at
+ * the start of the scan.  The first pass finds the lowest CCC to be
+ * emitted and stores it in u8c->nccc, the second pass emits the
+ * characters with this CCC and finds the next lowest CCC. This limits
+ * the number of passes to 1 + the number of different CCCs in the
+ * sequence being scanned.
+ *
+ * Therefore:
+ *  u8c->p  != NULL -> a decomposition is being scanned.
+ *  u8c->ss != NULL -> this is a repeating scan.
+ *  u8c->ccc == -1   -> this is the first scan of a repeating scan.
+ */
+int utf8byte(struct utf8cursor *u8c)
+{
+	utf8leaf_t *leaf;
+	int ccc;
+
+	for (;;) {
+		/* Check for the end of a decomposed character. */
+		if (u8c->p && *u8c->s == '\0') {
+			u8c->s = u8c->p;
+			u8c->p = NULL;
+		}
+
+		/* Check for end-of-string. */
+		if (!u8c->p && (u8c->len == 0 || *u8c->s == '\0')) {
+			/* There is no next byte. */
+			if (u8c->ccc == STOPPER)
+				return 0;
+			/* End-of-string during a scan counts as a stopper. */
+			ccc = STOPPER;
+			goto ccc_mismatch;
+		} else if ((*u8c->s & 0xC0) == 0x80) {
+			/* This is a continuation of the current character. */
+			if (!u8c->p)
+				u8c->len--;
+			return (unsigned char)*u8c->s++;
+		}
+
+		/* Look up the data for the current character. */
+		if (u8c->p) {
+			leaf = utf8lookup(u8c->um, u8c->n, u8c->hangul, u8c->s);
+		} else {
+			leaf = utf8nlookup(u8c->um, u8c->n, u8c->hangul,
+					   u8c->s, u8c->len);
+		}
+
+		/* No leaf found implies that the input is a binary blob. */
+		if (!leaf)
+			return -1;
+
+		ccc = LEAF_CCC(leaf);
+		/* Characters that are too new have CCC 0. */
+		if (u8c->um->tables->utf8agetab[LEAF_GEN(leaf)] >
+		    u8c->um->ntab[u8c->n]->maxage) {
+			ccc = STOPPER;
+		} else if (ccc == DECOMPOSE) {
+			u8c->len -= utf8clen(u8c->s);
+			u8c->p = u8c->s + utf8clen(u8c->s);
+			u8c->s = LEAF_STR(leaf);
+			/* Empty decomposition implies CCC 0. */
+			if (*u8c->s == '\0') {
+				if (u8c->ccc == STOPPER)
+					continue;
+				ccc = STOPPER;
+				goto ccc_mismatch;
+			}
+
+			leaf = utf8lookup(u8c->um, u8c->n, u8c->hangul, u8c->s);
+			if (!leaf)
+				return -1;
+			ccc = LEAF_CCC(leaf);
+		}
+
+		/*
+		 * If this is not a stopper, then see if it updates
+		 * the next canonical class to be emitted.
+		 */
+		if (ccc != STOPPER && u8c->ccc < ccc && ccc < u8c->nccc)
+			u8c->nccc = ccc;
+
+		/*
+		 * Return the current byte if this is the current
+		 * combining class.
+		 */
+		if (ccc == u8c->ccc) {
+			if (!u8c->p)
+				u8c->len--;
+			return (unsigned char)*u8c->s++;
+		}
+
+		/* Current combining class mismatch. */
+ccc_mismatch:
+		if (u8c->nccc == STOPPER) {
+			/*
+			 * Scan forward for the first canonical class
+			 * to be emitted.  Save the position from
+			 * which to restart.
+			 */
+			u8c->ccc = MINCCC - 1;
+			u8c->nccc = ccc;
+			u8c->sp = u8c->p;
+			u8c->ss = u8c->s;
+			u8c->slen = u8c->len;
+			if (!u8c->p)
+				u8c->len -= utf8clen(u8c->s);
+			u8c->s += utf8clen(u8c->s);
+		} else if (ccc != STOPPER) {
+			/* Not a stopper, and not the ccc we're emitting. */
+			if (!u8c->p)
+				u8c->len -= utf8clen(u8c->s);
+			u8c->s += utf8clen(u8c->s);
+		} else if (u8c->nccc != MAXCCC + 1) {
+			/* At a stopper, restart for next ccc. */
+			u8c->ccc = u8c->nccc;
+			u8c->nccc = MAXCCC + 1;
+			u8c->s = u8c->ss;
+			u8c->p = u8c->sp;
+			u8c->len = u8c->slen;
+		} else {
+			/* All done, proceed from here. */
+			u8c->ccc = STOPPER;
+			u8c->nccc = STOPPER;
+			u8c->sp = NULL;
+			u8c->ss = NULL;
+			u8c->slen = 0;
+		}
+	}
+}
+
+#if IS_MODULE(CONFIG_UNICODE_NORMALIZATION_KUNIT_TEST)
+EXPORT_SYMBOL_GPL(utf8version_is_supported);
+EXPORT_SYMBOL_GPL(utf8nlen);
+EXPORT_SYMBOL_GPL(utf8ncursor);
+EXPORT_SYMBOL_GPL(utf8byte);
+#endif

linux-7.2.3/fs/unicode/utf8n.h

@@ -0,0 +1,83 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2014 SGI.
+ * All rights reserved.
+ */
+
+#ifndef UTF8NORM_H
+#define UTF8NORM_H
+
+#include <linux/types.h>
+#include <linux/export.h>
+#include <linux/string.h>
+#include <linux/module.h>
+#include <linux/unicode.h>
+
+int utf8version_is_supported(const struct unicode_map *um, unsigned int version);
+
+/*
+ * Determine the length of the normalized from of the string,
+ * excluding any terminating NULL byte.
+ * Returns 0 if only ignorable code points are present.
+ * Returns -1 if the input is not valid UTF-8.
+ */
+ssize_t utf8nlen(const struct unicode_map *um, enum utf8_normalization n,
+		const char *s, size_t len);
+
+/* Needed in struct utf8cursor below. */
+#define UTF8HANGULLEAF	(12)
+
+/*
+ * Cursor structure used by the normalizer.
+ */
+struct utf8cursor {
+	const struct unicode_map *um;
+	enum utf8_normalization n;
+	const char	*s;
+	const char	*p;
+	const char	*ss;
+	const char	*sp;
+	unsigned int	len;
+	unsigned int	slen;
+	short int	ccc;
+	short int	nccc;
+	unsigned char	hangul[UTF8HANGULLEAF];
+};
+
+/*
+ * Initialize a utf8cursor to normalize a string.
+ * Returns 0 on success.
+ * Returns -1 on failure.
+ */
+int utf8ncursor(struct utf8cursor *u8c, const struct unicode_map *um,
+		enum utf8_normalization n, const char *s, size_t len);
+
+/*
+ * Get the next byte in the normalization.
+ * Returns a value > 0 && < 256 on success.
+ * Returns 0 when the end of the normalization is reached.
+ * Returns -1 if the string being normalized is not valid UTF-8.
+ */
+extern int utf8byte(struct utf8cursor *u8c);
+
+struct utf8data {
+	unsigned int maxage;
+	unsigned int offset;
+};
+
+struct utf8data_table {
+	const unsigned int *utf8agetab;
+	int utf8agetab_size;
+
+	const struct utf8data *utf8nfdicfdata;
+	int utf8nfdicfdata_size;
+
+	const struct utf8data *utf8nfdidata;
+	int utf8nfdidata_size;
+
+	const unsigned char *utf8data;
+};
+
+extern const struct utf8data_table utf8_data_table;
+
+#endif /* UTF8NORM_H */