diff --git a/NEWS b/NEWS
index 2a5d62337c..1a8782c01a 100644
--- a/NEWS
+++ b/NEWS
@@ -127,6 +127,10 @@ CHANGES WITH 240 in spe:
         * Support for disabling a particular cgroup controller within a sub-tree
           has been added through the DisableControllers= directive.
 
+        * cgroup_no_v1=all on the kernel command line now also implies
+          using the unified cgroup hierarchy, unless one explicitly passes
+          systemd.unified_cgroup_hierarchy=0 on the kernel command line.
+
         * The new "MemoryMin=" unit file property may now be used to set the
           memory usage protection limit of processes invoked by the unit. This
           controls the cgroupsv2 memory.min attribute. Similarly, the new
diff --git a/src/basic/cgroup-util.c b/src/basic/cgroup-util.c
index ec29a6f4c9..830a63c185 100644
--- a/src/basic/cgroup-util.c
+++ b/src/basic/cgroup-util.c
@@ -2706,6 +2706,7 @@ bool cg_is_unified_wanted(void) {
         int r;
         bool b;
         const bool is_default = DEFAULT_HIERARCHY == CGROUP_UNIFIED_ALL;
+        _cleanup_free_ char *c = NULL;
 
         /* If we have a cached value, return that. */
         if (wanted >= 0)
@@ -2716,11 +2717,19 @@ bool cg_is_unified_wanted(void) {
         if (cg_unified_flush() >= 0)
                 return (wanted = unified_cache >= CGROUP_UNIFIED_ALL);
 
-        /* Otherwise, let's see what the kernel command line has to say.
-         * Since checking is expensive, cache a non-error result. */
+        /* If we were explicitly passed systemd.unified_cgroup_hierarchy,
+         * respect that. */
         r = proc_cmdline_get_bool("systemd.unified_cgroup_hierarchy", &b);
+        if (r > 0)
+                return (wanted = b);
 
-        return (wanted = r > 0 ? b : is_default);
+        /* If we passed cgroup_no_v1=all with no other instructions, it seems
+         * highly unlikely that we want to use hybrid or legacy hierarchy. */
+        r = proc_cmdline_get_key("cgroup_no_v1", 0, &c);
+        if (r > 0 && streq_ptr(c, "all"))
+                return (wanted = true);
+
+        return (wanted = is_default);
 }
 
 bool cg_is_legacy_wanted(void) {
diff --git a/src/test/test-cgroup-util.c b/src/test/test-cgroup-util.c
index 058cc44c42..a3239d73f5 100644
--- a/src/test/test-cgroup-util.c
+++ b/src/test/test-cgroup-util.c
@@ -369,6 +369,17 @@ static void test_is_wanted(void) {
                          "systemd.unified_cgroup_hierarchy=0 "
                          "systemd.legacy_systemd_cgroup_controller=0", 1) >= 0);
         test_is_wanted_print(false);
+
+        /* cgroup_no_v1=all implies unified cgroup hierarchy, unless otherwise
+         * explicitly specified. */
+        assert_se(setenv("SYSTEMD_PROC_CMDLINE",
+                         "cgroup_no_v1=all", 1) >= 0);
+        test_is_wanted_print(false);
+
+        assert_se(setenv("SYSTEMD_PROC_CMDLINE",
+                         "cgroup_no_v1=all "
+                         "systemd.unified_cgroup_hierarchy=0", 1) >= 0);
+        test_is_wanted_print(false);
 }
 
 static void test_cg_tests(void) {