diff --git a/src/libnm-core-impl/tests/test-general.c b/src/libnm-core-impl/tests/test-general.c
index 4bbb3df451..30b1ca5da8 100644
--- a/src/libnm-core-impl/tests/test-general.c
+++ b/src/libnm-core-impl/tests/test-general.c
@@ -7,6 +7,7 @@
 
 #include "libnm-core-impl/nm-default-libnm-core.h"
 
+#include <ctype.h>
 #include <sys/types.h>
 #include <sys/stat.h>
 #include <fcntl.h>
@@ -87,21 +88,65 @@ G_STATIC_ASSERT(NM_UTILS_HWADDR_LEN_MAX == _NM_UTILS_HWADDR_LEN_MAX);
 static void
 test_nm_ascii_spaces(void)
 {
-    int               i;
-    const char *const S = NM_ASCII_SPACES;
+    int i;
 
-    for (i = 0; S[i]; i++)
-        g_assert(!strchr(&S[i + 1], S[i]));
+    {
+        const char *const S = NM_ASCII_SPACES;
 
-    for (i = 0; S[i] != '\0'; i++)
-        g_assert(g_ascii_isspace(S[i]));
+        for (i = 0; S[i]; i++)
+            g_assert(!strchr(&S[i + 1], S[i]));
 
-    g_assert(!g_ascii_isspace((char) 0));
-    for (i = 1; i < 0x100; i++) {
-        if (g_ascii_isspace((char) i))
-            g_assert(strchr(S, (char) i));
-        else
-            g_assert(!strchr(S, (char) i));
+        for (i = 0; S[i] != '\0'; i++)
+            g_assert(g_ascii_isspace(S[i]));
+
+        g_assert(!g_ascii_isspace((char) 0));
+        for (i = 1; i < 0x100; i++) {
+            g_assert((!!g_ascii_isspace((char) i)) == (!!strchr(S, (char) i)));
+        }
+    }
+
+    {
+        const char *const S = NM_ASCII_WHITESPACES;
+
+        for (i = 0; S[i]; i++)
+            g_assert(!strchr(&S[i + 1], S[i]));
+
+        for (i = 0; S[i] != '\0'; i++)
+            g_assert(nm_ascii_is_whitespace(S[i]));
+
+        g_assert(!nm_ascii_is_whitespace((char) 0));
+        for (i = 1; i < 0x100; i++) {
+            g_assert(nm_ascii_is_whitespace((char) i) == (!!strchr(S, (char) i)));
+        }
+    }
+
+    {
+        const char *const S = NM_ASCII_SPACES_CTYPE;
+
+        for (i = 0; S[i]; i++)
+            g_assert(!strchr(&S[i + 1], S[i]));
+
+        if (nm_streq0(g_getenv("LANG"), "C")) {
+            g_assert(!isspace((char) 0));
+            for (i = 1; i < 0x100; i++) {
+                g_assert((!!isspace((char) i)) == (!!strchr(S, (char) i)));
+            }
+        }
+    }
+
+    {
+        const char *const S = NM_ASCII_SPACES_KERNEL;
+
+        for (i = 0; S[i]; i++)
+            g_assert(!strchr(&S[i + 1], S[i]));
+
+        for (i = 0; S[i] != '\0'; i++)
+            g_assert(nm_ascii_is_space_kernel(S[i]));
+
+        g_assert(!nm_ascii_is_space_kernel((char) 0));
+        for (i = 1; i < 0x100; i++) {
+            g_assert(nm_ascii_is_space_kernel((char) i) == (!!strchr(S, (char) i)));
+        }
     }
 }
 
diff --git a/src/libnm-glib-aux/nm-macros-internal.h b/src/libnm-glib-aux/nm-macros-internal.h
index b74d732407..9972dc4452 100644
--- a/src/libnm-glib-aux/nm-macros-internal.h
+++ b/src/libnm-glib-aux/nm-macros-internal.h
@@ -1086,11 +1086,20 @@ nm_g_variant_equal(GVariant *a, GVariant *b)
 /* mirrors g_ascii_isspace() and what we consider spaces in general. */
 #define NM_ASCII_SPACES " \n\t\r\f"
 
-/* Like NM_ASCII_SPACES, but without "\f" (0x0c, Formfeed Page Break).
- * This is what for example systemd calls WHITESPACE and what it uses to tokenize
- * the kernel command line. */
+/* Like NM_ASCII_SPACES, but without "\f" (0x0c, Formfeed Page Break).  This is
+ * what for example systemd calls WHITESPACE and what it uses to tokenize the
+ * kernel command line. */
 #define NM_ASCII_WHITESPACES " \n\t\r"
 
+/* mirrors <ctype.h>'s isspace() with C locale. It's like NM_ASCII_SPACES but
+ * additionally also considers '\v' (vertical tab). */
+#define NM_ASCII_SPACES_CTYPE NM_ASCII_SPACES "\v"
+
+/* mirrors kernel's isspace() from "include/linux/ctype.h", which treats as
+ * space the common ASCII spaces, including '\v' (vertical tab), but also
+ * '\240' (non-breaking space, NBSP in Latin-1). */
+#define NM_ASCII_SPACES_KERNEL NM_ASCII_SPACES_CTYPE "\240"
+
 static inline gboolean
 nm_ascii_is_whitespace(char ch)
 {
@@ -1100,6 +1109,13 @@ nm_ascii_is_whitespace(char ch)
     return NM_IN_SET(ch, ' ', '\n', '\t', '\r');
 }
 
+static inline gboolean
+nm_ascii_is_space_kernel(char ch)
+{
+    /* Checks whether @ch is in NM_ASCII_SPACES_KERNEL. */
+    return NM_IN_SET(ch, ' ', '\n', '\t', '\r', '\f', '\v', '\240');
+}
+
 #define NM_ASCII_NEWLINE "\n\r"
 
 static inline gboolean
diff --git a/src/libnm-glib-aux/nm-shared-utils.c b/src/libnm-glib-aux/nm-shared-utils.c
index a51ff765d7..34a3af2053 100644
--- a/src/libnm-glib-aux/nm-shared-utils.c
+++ b/src/libnm-glib-aux/nm-shared-utils.c
@@ -5352,7 +5352,7 @@ nm_utils_ifname_valid_kernel(const char *name, GError **error)
 
         if (ch == '\0')
             return TRUE;
-        if (NM_IN_SET(ch, '/', ':') || g_ascii_isspace(ch)) {
+        if (NM_IN_SET(ch, '/', ':') || nm_ascii_is_space_kernel(ch)) {
             g_set_error_literal(error,
                                 NM_UTILS_ERROR,
                                 NM_UTILS_ERROR_UNKNOWN,