[RFC PATCH v2 18/25] sched/rt: Zero rt-cgroups default bandwidth

[Yuri Andriaccio <yurand2000@gmail.com>]

Set the default rt-cgroups runtime to zero, otherwise a cgroup-v1 kernel will
not be able to start SCHED_DEADLINE tasks. The bandwidth for rt-cgroups must
then be manually assigned after the kernel boots.

Allow zeroing the runtime of the root control group. This runtime only affects
the available bandwidth of the rt-cgroup hierarchy but not the SCHED_FIFO /
SCHED_RR tasks on the global runqueue.

Notes:
Disabling the root control group bandwidth should not cause any side effect, as
SCHED_FIFO / SCHED_RR tasks do not depend on it since the introduction of
fair_servers.

Signed-off-by: Yuri Andriaccio <yurand2000@gmail.com>
---
 kernel/sched/core.c     |  4 ++--
 kernel/sched/rt.c       | 13 +++++--------
 kernel/sched/syscalls.c |  2 +-
 3 files changed, 8 insertions(+), 11 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 85950e10bb1..3ac65c6af70 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -8719,7 +8719,7 @@ void __init sched_init(void)
 
 #ifdef CONFIG_RT_GROUP_SCHED
 	init_dl_bandwidth(&root_task_group.dl_bandwidth,
-			global_rt_period(), global_rt_runtime());
+			global_rt_period(), 0);
 #endif /* CONFIG_RT_GROUP_SCHED */
 
 #ifdef CONFIG_CGROUP_SCHED
@@ -9348,7 +9348,7 @@ static int cpu_cgroup_can_attach(struct cgroup_taskset *tset)
 		goto scx_check;
 
 	cgroup_taskset_for_each(task, css, tset) {
-		if (!sched_rt_can_attach(css_tg(css), task))
+		if (rt_task(task) && !sched_rt_can_attach(css_tg(css), task))
 			return -EINVAL;
 	}
 scx_check:
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index f37ac9100d1..75a6860c2e2 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -2122,13 +2122,6 @@ static int tg_set_rt_bandwidth(struct task_group *tg,
 	static DEFINE_MUTEX(rt_constraints_mutex);
 	int i, err = 0;
 
-	/*
-	 * Disallowing the root group RT runtime is BAD, it would disallow the
-	 * kernel creating (and or operating) RT threads.
-	 */
-	if (tg == &root_task_group && rt_runtime == 0)
-		return -EINVAL;
-
 	/* No period doesn't make any sense. */
 	if (rt_period == 0)
 		return -EINVAL;
@@ -2215,8 +2208,12 @@ static int sched_rt_global_constraints(void)
 
 int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk)
 {
+	/* Allow executing in the root cgroup regardless of allowed bandwidth */
+	if (tg == &root_task_group)
+		return 1;
+
 	/* Don't accept real-time tasks when there is no way for them to run */
-	if (rt_group_sched_enabled() && rt_task(tsk) && tg->dl_bandwidth.dl_runtime == 0)
+	if (rt_group_sched_enabled() && tg->dl_bandwidth.dl_runtime == 0)
 		return 0;
 
 	return 1;
diff --git a/kernel/sched/syscalls.c b/kernel/sched/syscalls.c
index 7c1f7649477..71f20be6f29 100644
--- a/kernel/sched/syscalls.c
+++ b/kernel/sched/syscalls.c
@@ -633,7 +633,7 @@ int __sched_setscheduler(struct task_struct *p,
 		 */
 		if (rt_group_sched_enabled() &&
 				dl_bandwidth_enabled() && rt_policy(policy) &&
-				task_group(p)->dl_bandwidth.dl_runtime == 0 &&
+				!sched_rt_can_attach(task_group(p), p) &&
 				!task_group_is_autogroup(task_group(p))) {
 			retval = -EPERM;
 			goto unlock;
-- 
2.50.1