fs: support non-ASCII VFAT labels via locale-derived codepage

src/plugins/fs/vfat.c

@@ -276,43 +276,147 @@ gboolean bd_fs_vfat_repair (const gchar *device, const BDExtraArg **extra, GErro
     return ret;
 }
 
+/**
+ * _vfat_locale_codepage:
+ *
+ * Determine the DOS/OEM codepage that matches the system locale, mirroring
+ * how Windows derives the OEM codepage from the system locale. FAT volume
+ * labels are stored as 8-bit OEM codepage bytes, so the codepage used to
+ * write a label must match the locale of the system that created it.
+ *
+ * The locale is read from the environment (LC_ALL > LC_CTYPE > LANG) rather
+ * than calling setlocale(), to avoid side effects on the host process
+ * (libblockdev runs inside udisksd).
+ *
+ * Returns: the codepage number, or 850 (the dosfstools default) if the locale
+ *          cannot be determined or has no known mapping.
+ */
+static guint
+_vfat_locale_codepage (void) {
+    const gchar *locale = g_getenv ("LC_ALL");
+    if (locale == NULL || *locale == '\0')
+        locale = g_getenv ("LC_CTYPE");
+    if (locale == NULL || *locale == '\0')
+        locale = g_getenv ("LANG");
+
+    if (locale == NULL || *locale == '\0' ||
+        g_strcmp0 (locale, "C") == 0 || g_strcmp0 (locale, "POSIX") == 0 ||
+        g_str_has_prefix (locale, "C."))
+        return 850;
+
+    /* Extract the language[_territory] part before '.' or '@' */
+    g_autofree gchar *lang = g_strdup (locale);
+    gchar *p;
+    p = strchr (lang, '.');
+    if (p) *p = '\0';
+    p = strchr (lang, '@');
+    if (p) *p = '\0';
+
+    if (g_str_has_prefix (lang, "zh_CN") || g_str_has_prefix (lang, "zh_SG"))
+        return 936;   /* GBK */
+    if (g_str_has_prefix (lang, "zh_TW") || g_str_has_prefix (lang, "zh_HK"))
+        return 950;   /* Big5 */
+    if (g_str_has_prefix (lang, "ja"))
+        return 932;   /* Shift-JIS */
+    if (g_str_has_prefix (lang, "ko"))
+        return 949;   /* UHC (Korean) */
+    if (g_str_has_prefix (lang, "ru") || g_str_has_prefix (lang, "uk") ||
+        g_str_has_prefix (lang, "be"))
+        return 866;   /* Cyrillic (DOS) */
+
+    return 850;
+}
+
+/**
+ * _vfat_label_from_codepage:
+ * @label: (nullable): raw OEM codepage bytes from the filesystem
+ *
+ * Convert a label read from a VFAT filesystem (raw OEM codepage bytes, as
+ * returned by blkid) to UTF-8, using the locale-derived codepage. Pure
+ * ASCII labels and the default codepage (850) need no conversion.
+ *
+ * Returns: (transfer full): the UTF-8 label, or the original bytes on
+ *          conversion failure. Never %NULL.
+ */
+static gchar *
+_vfat_label_from_codepage (const gchar *label) {
+    if (label == NULL || *label == '\0')
+        return g_strdup (label ? label : "");
+
+    if (g_str_is_ascii (label))
+        return g_strdup (label);
+
+    guint cp = _vfat_locale_codepage ();
+    gchar cp_name[16];
+    g_snprintf (cp_name, sizeof (cp_name), "CP%u", cp);
+
+    GError *conv_error = NULL;
+    gchar *utf8 = g_convert (label, -1, "UTF-8", cp_name, NULL, NULL, &conv_error);
+    if (utf8 != NULL)
+        return utf8;
+
+    /* Conversion failed; return a valid UTF-8 string instead of raw bytes
+     * that may not be valid UTF-8. */
+    g_error_free (conv_error);
+    return g_utf8_make_valid (label, -1);
+}
+
 /**
  * bd_fs_vfat_set_label:
  * @device: the device containing the file system to set label for
  * @label: label to set
  * @error: (out) (optional): place to store error (if any)
  *
- * Returns: whether the label of vfat file system on the @device was
- *          successfully set or not
+ * Sets the label of a VFAT filesystem. For labels containing non-ASCII
+ * characters, the OEM codepage matching the system locale is passed to
+ * fatlabel via the -c option, so that labels in CJK, Cyrillic and other
+ * non-Western-European scripts are encoded correctly.
+ *
+ * Returns: whether the label was successfully set or not
  *
  * Tech category: %BD_FS_TECH_VFAT-%BD_FS_TECH_MODE_SET_LABEL
  */
 gboolean bd_fs_vfat_set_label (const gchar *device, const gchar *label, GError **error) {
-    const gchar *args[4] = {"fatlabel", device, NULL, NULL};
+    /* "fatlabel" "-c" "<cp>" <device> <label> -- at most 5 non-NULL entries */
+    const gchar *args[6] = {"fatlabel", NULL, NULL, NULL, NULL, NULL};
     UtilDep dep = {"fatlabel", "4.2", "--version", "fatlabel\\s+([\\d\\.]+).+"};
     gchar *label_up = NULL;
+    gchar *codepage_str = NULL;
     gboolean new_vfat = FALSE;
     gboolean ret;
+    gint idx = 1;
 
     if (!check_deps (&avail_deps, DEPS_FATLABEL_MASK, deps, DEPS_LAST, &deps_check_lock, error))
         return FALSE;
 
+    /* For labels with non-ASCII characters, pass the OEM codepage matching
+     * the system locale to fatlabel so it can encode them correctly. */
+    if (label && *label && !g_str_is_ascii (label)) {
+        guint cp = _vfat_locale_codepage ();
+        codepage_str = g_strdup_printf ("%u", cp);
+        args[idx++] = "-c";
+        args[idx++] = codepage_str;
+    }
+
+    args[idx++] = device;
+
     if (!label || g_strcmp0 (label, "") == 0) {
         /* fatlabel >= 4.2 refuses to set empty label */
         new_vfat = bd_utils_check_util_version (dep.name, dep.version,
                                                 dep.ver_arg, dep.ver_regexp,
                                                 NULL);
         if (new_vfat)
-            args[2] = "--reset";
-    }
-
-    /* forcefully convert the label uppercase in case no reset was requested */
-    if (label && args[2] == NULL) {
+            args[idx++] = "--reset";
+    } else {
+        /* forcefully convert the label uppercase; g_ascii_strup only
+         * uppercases a-z, leaving multi-byte UTF-8 sequences unchanged */
         label_up = g_ascii_strup (label, -1);
-        args[2] = label_up;
+        args[idx++] = label_up;
     }
+
     ret = bd_utils_exec_and_report_error (args, NULL, error);
     g_free (label_up);
+    g_free (codepage_str);
 
     return ret;
 }
@@ -331,9 +435,36 @@ gboolean bd_fs_vfat_check_label (const gchar *label, GError **error) {
     const gchar *forbidden = "\"*/:<>?\\|";
     guint n;
 
-    if (strlen (label) > 11) {
+    if (label == NULL) {
         g_set_error_literal (error, BD_FS_ERROR, BD_FS_ERROR_LABEL_INVALID,
-                             "Label for VFAT filesystem must be at most 11 characters long.");
+                             "Label cannot be NULL.");
+        return FALSE;
+    }
+
+    /* VFAT labels are stored as 11 raw OEM codepage bytes. For non-ASCII
+     * labels, convert to the locale-derived OEM codepage and check the
+     * resulting byte count so that check_label matches what set_label
+     * (and fatlabel) actually accept. */
+    gsize label_len = strlen (label);
+    if (!g_str_is_ascii (label)) {
+        gchar cp_name[16];
+        g_snprintf (cp_name, sizeof (cp_name), "CP%u", _vfat_locale_codepage ());
+        g_autofree gchar *oem = NULL;
+        gsize written = 0;
+        GError *conv_error = NULL;
+        oem = g_convert (label, -1, cp_name, "UTF-8", NULL, &written, &conv_error);
+        if (oem == NULL) {
+            g_error_free (conv_error);
+            g_set_error_literal (error, BD_FS_ERROR, BD_FS_ERROR_LABEL_INVALID,
+                                 "Label cannot be encoded in the OEM codepage of the current locale.");
+            return FALSE;
+        }
+        label_len = written;
+    }
+
+    if (label_len > 11) {
+        g_set_error_literal (error, BD_FS_ERROR, BD_FS_ERROR_LABEL_INVALID,
+                             "Label for VFAT filesystem must be at most 11 bytes long.");
         return FALSE;
     }
 
@@ -454,6 +585,14 @@ BDFSVfatInfo* bd_fs_vfat_get_info (const gchar *device, GError **error) {
         return NULL;
     }
 
+    /* blkid returns raw OEM codepage bytes; convert to UTF-8 so the label
+     * round-trips with bd_fs_vfat_set_label(). */
+    {
+        gchar *utf8_label = _vfat_label_from_codepage (ret->label);
+        g_free (ret->label);
+        ret->label = utf8_label;
+    }
+
     success = bd_utils_exec_and_capture_output (args, NULL, &output, error);
     if (!success) {
         /* error is already populated */

tests/fs_tests/vfat_test.py

@@ -1,5 +1,8 @@
+import os
 import re
+import subprocess
 import tempfile
+import unittest
 
 from packaging.version import Version
 
@@ -23,6 +26,15 @@ def _get_dosfstools_version():
 DOSFSTOOLS_VERSION = _get_dosfstools_version()
 
 
+def _has_codepage(cp):
+    """Check whether the given DOS codepage is available via iconv."""
+    try:
+        p = subprocess.run(["iconv", "-l"], capture_output=True, text=True)
+        return "CP%d" % cp in p.stdout or "CP%d//" % cp in p.stdout
+    except Exception:
+        return False
+
+
 class VfatNoDevTestCase(FSNoDevTestCase):
     pass
 
@@ -228,7 +240,7 @@ def test_vfat_set_label(self):
         succ = BlockDev.fs_vfat_check_label("TEST_LABEL")
         self.assertTrue(succ)
 
-        with self.assertRaisesRegex(GLib.GError, "at most 11 characters long."):
+        with self.assertRaisesRegex(GLib.GError, "at most 11 bytes long."):
             BlockDev.fs_vfat_check_label(12 * "a")
 
 
@@ -281,6 +293,106 @@ def test_vfat_set_uuid(self):
             BlockDev.fs_vfat_check_uuid(10 * "f")
 
 
+class VfatCheckLabel(VfatNoDevTestCase):
+    """Tests for bd_fs_vfat_check_label (no device needed)."""
+
+    def test_vfat_check_label_byte_limit(self):
+        """The limit is 11 bytes, not 11 characters.
+
+        A 4-character CJK label (12 UTF-8 bytes) exceeds the 11-byte limit
+        and must be rejected, even though it is only 4 characters long.
+        """
+        succ = BlockDev.fs_vfat_check_label("a" * 11)
+        self.assertTrue(succ)
+
+        with self.assertRaisesRegex(GLib.GError, "at most 11 bytes long."):
+            BlockDev.fs_vfat_check_label("a" * 12)
+
+    def test_vfat_check_label_non_ascii_byte_count(self):
+        """Non-ASCII labels are checked by byte count, not character count."""
+        # 3 CJK chars + 1 ASCII char = 10 UTF-8 bytes <= 11
+        succ = BlockDev.fs_vfat_check_label("测试U盘")
+        self.assertTrue(succ)
+
+        # 4 CJK chars = 12 UTF-8 bytes > 11
+        with self.assertRaisesRegex(GLib.GError, "at most 11 bytes long."):
+            BlockDev.fs_vfat_check_label("测试优盘")
+
+    def test_vfat_check_label_forbidden_chars(self):
+        """Forbidden characters must still be rejected."""
+        for ch in '"*/:<>?\\|':
+            with self.assertRaisesRegex(GLib.GError, "not supported in VFAT labels"):
+                BlockDev.fs_vfat_check_label("A" + ch + "B")
+
+
+@unittest.skipUnless(_has_codepage(936),
+                    "DOS codepage 936 (GBK) not available via iconv")
+class VfatSetLabelNonAscii(VfatTestCase):
+    """Tests for non-ASCII (CJK) VFAT labels using locale-derived codepage."""
+
+    @classmethod
+    def setUpClass(cls):
+        super(VfatSetLabelNonAscii, cls).setUpClass()
+        cls._saved_env = {}
+        for var in ("LC_ALL", "LC_CTYPE", "LANG"):
+            cls._saved_env[var] = os.environ.get(var)
+
+    @classmethod
+    def tearDownClass(cls):
+        for var, val in cls._saved_env.items():
+            if val is None:
+                os.environ.pop(var, None)
+            else:
+                os.environ[var] = val
+        super(VfatSetLabelNonAscii, cls).tearDownClass()
+
+    def setUp(self):
+        super(VfatSetLabelNonAscii, self).setUp()
+        # _vfat_locale_codepage reads LC_ALL > LC_CTYPE > LANG from the
+        # environment; set a CJK locale so it maps to codepage 936 (GBK).
+        os.environ["LC_ALL"] = "zh_CN.UTF-8"
+
+    def test_vfat_set_non_ascii_label_roundtrip(self):
+        """Set a CJK label and verify get_info reads it back as UTF-8."""
+
+        succ = BlockDev.fs_vfat_mkfs(self.loop_devs[0], self._mkfs_options)
+        self.assertTrue(succ)
+
+        label = "测试U盘"  # 4 CJK/Latin chars, 7 GBK bytes, 10 UTF-8 bytes
+        succ = BlockDev.fs_vfat_set_label(self.loop_devs[0], label)
+        self.assertTrue(succ)
+
+        fi = BlockDev.fs_vfat_get_info(self.loop_devs[0])
+        self.assertTrue(fi)
+        self.assertEqual(fi.label, label)
+
+    def test_vfat_set_ascii_label_unchanged(self):
+        """Pure-ASCII labels must not be affected by the codepage change."""
+
+        succ = BlockDev.fs_vfat_mkfs(self.loop_devs[0], self._mkfs_options)
+        self.assertTrue(succ)
+
+        succ = BlockDev.fs_vfat_set_label(self.loop_devs[0], "HELLO")
+        self.assertTrue(succ)
+
+        fi = BlockDev.fs_vfat_get_info(self.loop_devs[0])
+        self.assertTrue(fi)
+        self.assertEqual(fi.label, "HELLO")
+
+    def test_vfat_set_label_max_cjk(self):
+        """5 CJK characters (10 GBK bytes) fit in 11 DOS bytes; 6 do not."""
+
+        succ = BlockDev.fs_vfat_mkfs(self.loop_devs[0], self._mkfs_options)
+        self.assertTrue(succ)
+
+        # 5 CJK chars = 10 GBK bytes <= 11 -- should succeed
+        five = "测" * 5
+        succ = BlockDev.fs_vfat_set_label(self.loop_devs[0], five)
+        self.assertTrue(succ)
+        fi = BlockDev.fs_vfat_get_info(self.loop_devs[0])
+        self.assertEqual(fi.label, five)
+
+
 @utils.required_plugins(("tools",))
 class VfatResize(VfatTestCase):
     def test_vfat_resize(self):