path: root/0014-add-CONFIG_ECHO_SCHED.patch

From d37ebcbf15c162531a18b7a4b131f908c2be39ca Mon Sep 17 00:00:00 2001
From: hamadmarri <hamad.s.almarri@gmail.com>
Date: Mon, 25 Mar 2024 23:36:53 +0300
Subject: [PATCH 14/16] add CONFIG_ECHO_SCHED

---
 include/linux/sched.h  |  9 ++++++---
 init/Kconfig           | 10 +++++++++-
 kernel/Kconfig.preempt |  2 +-
 kernel/sched/Makefile  |  4 ++++
 kernel/sched/core.c    | 14 ++++++++++++++
 kernel/sched/debug.c   |  2 ++
 kernel/sched/idle.c    | 12 +++++++++---
 kernel/sched/sched.h   | 10 ++++++++++
 8 files changed, 55 insertions(+), 8 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 5b22c1f26..947d94be2 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -533,6 +533,7 @@ struct sched_statistics {
 #endif /* CONFIG_SCHEDSTATS */
 } ____cacheline_aligned;
 
+#ifdef CONFIG_ECHO_SCHED
 struct bs_node {
 	struct bs_node*                 next;
 	u64				c_vrt_start;
@@ -543,6 +544,7 @@ struct bs_node {
 #endif
 	u64				est;
 };
+#endif
 
 struct sched_entity {
 	/* For load-balancing: */
@@ -553,17 +555,18 @@ struct sched_entity {
 
 	struct list_head		group_node;
 	unsigned int			on_rq;
-
+#ifdef CONFIG_ECHO_SCHED
 	struct bs_node                  bs_node;
-
+#endif
 	u64				exec_start;
 	u64				sum_exec_runtime;
 	u64				prev_sum_exec_runtime;
 	u64				vruntime;
 	s64				vlag;
 	u64				slice;
+#ifdef CONFIG_ECHO_SCHED
 	bool				yielded;
-
+#endif
 	u64				nr_migrations;
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
diff --git a/init/Kconfig b/init/Kconfig
index a8bf21f68..933ec5c9a 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -130,6 +130,12 @@ config THREAD_INFO_IN_TASK
 	  One subtle change that will be needed is to use try_get_task_stack()
 	  and put_task_stack() in save_thread_stack_tsk() and get_wchan().
 
+config ECHO_SCHED
+	bool "ECHO CPU Scheduler"
+	default y
+	help
+	  https://github.com/hamadmarri/ECHO-CPU-Scheduler
+
 menu "General setup"
 
 config BROKEN
@@ -1008,11 +1014,12 @@ menuconfig CGROUP_SCHED
 if CGROUP_SCHED
 config FAIR_GROUP_SCHED
 	bool "Group scheduling for SCHED_OTHER"
-	depends on CGROUP_SCHED
+	depends on CGROUP_SCHED && !ECHO_SCHED
 	default n
 
 config CFS_BANDWIDTH
 	bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
+	depends on !ECHO_SCHED
 	depends on FAIR_GROUP_SCHED
 	default n
 	help
@@ -1281,6 +1288,7 @@ config CHECKPOINT_RESTORE
 
 config SCHED_AUTOGROUP
 	bool "Automatic process group scheduling"
+	depends on !ECHO_SCHED
 	select CGROUPS
 	select CGROUP_SCHED
 	select FAIR_GROUP_SCHED
diff --git a/kernel/Kconfig.preempt b/kernel/Kconfig.preempt
index 7d671d4e6..d54bb52cc 100644
--- a/kernel/Kconfig.preempt
+++ b/kernel/Kconfig.preempt
@@ -117,7 +117,7 @@ config PREEMPT_DYNAMIC
 
 config SCHED_CORE
 	bool "Core Scheduling for SMT"
-	depends on SCHED_SMT
+	depends on SCHED_SMT && !ECHO_SCHED
 	default n
 	help
 	  This option permits Core Scheduling, a means of coordinated task
diff --git a/kernel/sched/Makefile b/kernel/sched/Makefile
index f550be3a2..f78ee1bff 100644
--- a/kernel/sched/Makefile
+++ b/kernel/sched/Makefile
@@ -29,6 +29,10 @@ endif
 # build parallelizes well and finishes roughly at once:
 #
 obj-y += core.o
+ifeq ($(CONFIG_ECHO_SCHED),y)
 obj-y += bs.o
+else
+obj-y += fair.o
+endif
 obj-y += build_policy.o
 obj-y += build_utility.o
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 83fa2c8c2..ff6e2d0a1 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -3347,6 +3347,7 @@ void relax_compatible_cpus_allowed_ptr(struct task_struct *p)
 	WARN_ON_ONCE(ret);
 }
 
+#ifdef CONFIG_ECHO_SCHED
 inline void inc_nr_lat_sensitive(unsigned int cpu, struct task_struct *p)
 {
 	if (per_cpu(nr_lat_sensitive, cpu) == 0 || per_cpu(nr_lat_sensitive, cpu) == -10)
@@ -3362,6 +3363,7 @@ inline void dec_nr_lat_sensitive(unsigned int cpu)
 			per_cpu(nr_lat_sensitive, cpu) = -1;
 	}
 }
+#endif
 
 void set_task_cpu(struct task_struct *p, unsigned int new_cpu)
 {
@@ -4542,8 +4544,10 @@ static void __sched_fork(unsigned long clone_flags, struct task_struct *p)
 	p->se.vlag			= 0;
 	p->se.slice			= sysctl_sched_base_slice;
 
+#ifdef CONFIG_ECHO_SCHED
 	p->se.bs_node.vburst		= 0;
 	p->se.bs_node.est		= 0;
+#endif
 
 	INIT_LIST_HEAD(&p->se.group_node);
 
@@ -4707,6 +4711,7 @@ static int sysctl_schedstats(struct ctl_table *table, int write, void *buffer,
 
 #ifdef CONFIG_SYSCTL
 static struct ctl_table sched_core_sysctls[] = {
+#ifdef CONFIG_ECHO_SCHED
 	{
 		.procname	= "sched_bs_shared_quota",
 		.data		= &bs_shared_quota,
@@ -4714,6 +4719,7 @@ static struct ctl_table sched_core_sysctls[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec,
 	},
+#endif
 #ifdef CONFIG_SCHEDSTATS
 	{
 		.procname       = "sched_schedstats",
@@ -5716,10 +5722,12 @@ void scheduler_tick(void)
 	if (curr->flags & PF_WQ_WORKER)
 		wq_worker_tick(curr);
 
+#ifdef CONFIG_ECHO_SCHED
 	if (idle_cpu(cpu))
 		inc_nr_lat_sensitive(cpu, NULL);
 	else
 		dec_nr_lat_sensitive(cpu);
+#endif
 
 #ifdef CONFIG_SMP
 	rq->idle_balance = idle_cpu(cpu);
@@ -9929,7 +9937,9 @@ LIST_HEAD(task_groups);
 static struct kmem_cache *task_group_cache __ro_after_init;
 #endif
 
+#ifdef CONFIG_ECHO_SCHED
 DEFINE_PER_CPU(int, nr_lat_sensitive);
+#endif
 
 void __init sched_init(void)
 {
@@ -9946,7 +9956,9 @@ void __init sched_init(void)
 
 	wait_bit_init();
 
+#ifdef CONFIG_ECHO_SCHED
 	printk(KERN_INFO "ECHO CPU scheduler v6.7 by Hamad Al Marri.");
+#endif
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 	ptr += 2 * nr_cpu_ids * sizeof(void **);
@@ -10067,7 +10079,9 @@ void __init sched_init(void)
 		hrtick_rq_init(rq);
 		atomic_set(&rq->nr_iowait, 0);
 
+#ifdef CONFIG_ECHO_SCHED
 		per_cpu(nr_lat_sensitive, i) = 0;
+#endif
 
 #ifdef CONFIG_SCHED_CORE
 		rq->core = rq;
diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 370dfb693..ec7d41bc6 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -1003,9 +1003,11 @@ void proc_sched_show_task(struct task_struct *p, struct pid_namespace *ns,
 	PN(se.exec_start);
 	PN(se.vruntime);
 	PN(se.sum_exec_runtime);
+#ifdef CONFIG_ECHO_SCHED
 	PN(se.bs_node.vburst);
 	PN(se.bs_node.prev_vburst);
 	PN(se.bs_node.est);
+#endif
 
 	nr_switches = p->nvcsw + p->nivcsw;
 
diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c
index c8f92fefd..dbfc30710 100644
--- a/kernel/sched/idle.c
+++ b/kernel/sched/idle.c
@@ -237,8 +237,9 @@ static void cpuidle_idle_call(void)
 static void do_idle(void)
 {
 	int cpu = smp_processor_id();
+#ifdef CONFIG_ECHO_SCHED
 	int pm_disabled = per_cpu(nr_lat_sensitive, cpu);
-
+#endif
 	/*
 	 * Check if we need to update blocked load
 	 */
@@ -306,7 +307,11 @@ static void do_idle(void)
 		 * broadcast device expired for us, we don't want to go deep
 		 * idle as we know that the IPI is going to arrive right away.
 		 */
-		if (pm_disabled > 0 || cpu_idle_force_poll || tick_check_broadcast_expired()) {
+		if (
+#ifdef CONFIG_ECHO_SCHED
+			pm_disabled > 0 ||
+#endif
+			cpu_idle_force_poll || tick_check_broadcast_expired()) {
 			tick_nohz_idle_restart_tick();
 			cpu_idle_poll();
 			dec_nr_lat_sensitive(cpu);
@@ -314,9 +319,10 @@ static void do_idle(void)
 			cpuidle_idle_call();
 		}
 
+#ifdef CONFIG_ECHO_SCHED
 		if (pm_disabled < 0)
 			dec_nr_lat_sensitive(cpu);
-
+#endif
 		arch_cpu_idle_exit();
 	}
 
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index d8758c2ce..56b5c0114 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -109,7 +109,9 @@ extern int sysctl_sched_rt_period;
 extern int sysctl_sched_rt_runtime;
 extern int sched_rr_timeslice;
 
+#ifdef CONFIG_ECHO_SCHED
 extern unsigned int bs_shared_quota;
+#endif
 
 /*
  * Helpers for converting nanosecond timing to jiffy resolution
@@ -576,7 +578,9 @@ struct cfs_rq {
 	unsigned int		h_nr_running;      /* SCHED_{NORMAL,BATCH,IDLE} */
 	unsigned int		idle_nr_running;   /* SCHED_IDLE */
 	unsigned int		idle_h_nr_running; /* SCHED_IDLE */
+#ifdef CONFIG_ECHO_SCHED
 	u64			local_cand_est;
+#endif
 	s64			avg_vruntime;
 	u64			avg_load;
 
@@ -598,8 +602,10 @@ struct cfs_rq {
 	 * It is set to NULL otherwise (i.e when none are currently running).
 	 */
 	struct sched_entity	*curr;
+#ifdef CONFIG_ECHO_SCHED
 	struct bs_node		*head;
 	struct bs_node		*q2_head;
+#endif
 	struct sched_entity	*next;
 
 #ifdef	CONFIG_SCHED_DEBUG
@@ -1895,7 +1901,9 @@ DECLARE_PER_CPU(struct sched_domain_shared __rcu *, sd_llc_shared);
 DECLARE_PER_CPU(struct sched_domain __rcu *, sd_numa);
 DECLARE_PER_CPU(struct sched_domain __rcu *, sd_asym_packing);
 DECLARE_PER_CPU(struct sched_domain __rcu *, sd_asym_cpucapacity);
+#ifdef CONFIG_ECHO_SCHED
 DECLARE_PER_CPU(int, nr_lat_sensitive);
+#endif
 extern struct static_key_false sched_asym_cpucapacity;
 extern struct static_key_false sched_cluster_active;
 
@@ -2553,8 +2561,10 @@ extern void wakeup_preempt(struct rq *rq, struct task_struct *p, int flags);
 #define SCHED_NR_MIGRATE_BREAK 32
 #endif
 
+#ifdef CONFIG_ECHO_SCHED
 extern inline void inc_nr_lat_sensitive(unsigned int cpu, struct task_struct *p);
 extern inline void dec_nr_lat_sensitive(unsigned int cpu);
+#endif
 
 extern const_debug unsigned int sysctl_sched_nr_migrate;
 extern const_debug unsigned int sysctl_sched_migration_cost;
-- 
2.44.0