vruntime.patch

diff --git a/arch/x86/entry/syscalls/syscall_32.tbl b/arch/x86/entry/syscalls/syscall_32.tbl
index 0985d8333..97b6336a0 100644
--- a/arch/x86/entry/syscalls/syscall_32.tbl
+++ b/arch/x86/entry/syscalls/syscall_32.tbl
@@ -453,3 +453,4 @@
 446	i386	landlock_restrict_self	sys_landlock_restrict_self
 447	i386	memfd_secret		sys_memfd_secret
 448	i386	process_mrelease	sys_process_mrelease
+449 i386  set_process_inactive  sys_set_process_inactive
diff --git a/arch/x86/entry/syscalls/syscall_64.tbl b/arch/x86/entry/syscalls/syscall_64.tbl
index 18b5500ea..898e08665 100644
--- a/arch/x86/entry/syscalls/syscall_64.tbl
+++ b/arch/x86/entry/syscalls/syscall_64.tbl
@@ -370,6 +370,7 @@
 446	common	landlock_restrict_self	sys_landlock_restrict_self
 447	common	memfd_secret		sys_memfd_secret
 448	common	process_mrelease	sys_process_mrelease
+449	common	set_process_inactive  sys_set_process_inactive
 
 #
 # Due to a historical design error, certain syscalls are numbered differently
diff --git a/benchmark.c b/benchmark.c
new file mode 100644
index 000000000..04b2750e5
--- /dev/null
+++ b/benchmark.c
@@ -0,0 +1,42 @@
+#define _GNU_SOURCE
+#include <pthread.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/time.h>
+#include <sys/syscall.h>
+#include <errno.h>
+
+#define SYS_set_process_inactive 449
+
+void *worker(void *arg) {
+  while(1);
+  return NULL;
+}
+
+int main(void) {
+  unsigned long penalty = 1000000;
+  int ret = syscall(SYS_set_process_inactive, penalty);
+  if (ret < 0) {
+    printf("syscall failed\n");
+    return 0;
+  }
+  int num_workers = sysconf(_SC_NPROCESSORS_ONLN);
+  if (num_workers < 1)
+    return 1;
+
+  struct timeval start, end;
+  double elapsed;
+
+  pthread_t *threads = malloc(num_workers * sizeof(pthread_t));
+
+  for (int i = 0; i < num_workers; i++) {
+    if (pthread_create(&threads[i], NULL, worker, NULL) != 0)
+      return 1;
+  }
+  for (int i = 0; i < num_workers; i++)
+    pthread_join(threads[i], NULL);
+  free(threads);
+  return 0;
+}
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index b8037a46f..296111c6f 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -313,6 +313,7 @@ static inline void addr_limit_user_check(void)
  * include the prototypes if CONFIG_ARCH_HAS_SYSCALL_WRAPPER is enabled.
  */
 #ifndef CONFIG_ARCH_HAS_SYSCALL_WRAPPER
+asmlinkage long sys_set_process_inactive(u64 penalty);
 asmlinkage long sys_io_setup(unsigned nr_reqs, aio_context_t __user *ctx);
 asmlinkage long sys_io_destroy(aio_context_t ctx);
 asmlinkage long sys_io_submit(aio_context_t, long,
diff --git a/include/uapi/asm-generic/unistd.h b/include/uapi/asm-generic/unistd.h
index 17a5a317f..4fb330132 100644
--- a/include/uapi/asm-generic/unistd.h
+++ b/include/uapi/asm-generic/unistd.h
@@ -880,8 +880,11 @@ __SYSCALL(__NR_memfd_secret, sys_memfd_secret)
 #define __NR_process_mrelease 448
 __SYSCALL(__NR_process_mrelease, sys_process_mrelease)
 
+#define __NR_set_process_inactive 449
+__SYSCALL(__NR_set_process_inactive, sys_set_process_inactive)
+
 #undef __NR_syscalls
-#define __NR_syscalls 449
+#define __NR_syscalls 450
 
 /*
  * 32 bit systems traditionally used different
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 74ef3d8f4..c51265336 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -45,6 +45,9 @@ EXPORT_TRACEPOINT_SYMBOL_GPL(sched_update_nr_running_tp);
 
 DEFINE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues);
 
+LIST_HEAD(inactive_tgid_list);
+DEFINE_SPINLOCK(inactive_tgid_lock);
+
 #ifdef CONFIG_SCHED_DEBUG
 /*
  * Debugging: various feature bits
@@ -7263,6 +7266,32 @@ static struct task_struct *find_process_by_pid(pid_t pid)
 	return pid ? find_task_by_vpid(pid) : current;
 }
 
+SYSCALL_DEFINE1(set_process_inactive, u64, penalty)
+{
+	struct inactive_tgid *entry;
+	unsigned long flags;
+	spin_lock_irqsave(&inactive_tgid_lock, flags);
+
+	list_for_each_entry(entry, &inactive_tgid_list, list) {
+		if (entry->tgid == current->tgid) {
+			spin_unlock_irqrestore(&inactive_tgid_lock, flags);
+			return 0;
+		}
+	}
+	entry = kmalloc(sizeof(struct inactive_tgid), GFP_ATOMIC);
+	if (!entry) {
+			spin_unlock_irqrestore(&inactive_tgid_lock, flags);
+			return -ENOMEM;
+	}
+	entry->tgid = current->tgid;
+	entry->penalty = penalty;
+	INIT_LIST_HEAD(&entry->list);
+	list_add(&entry->list, &inactive_tgid_list);
+
+	spin_unlock_irqrestore(&inactive_tgid_lock, flags);
+	return 0;
+}
+
 /*
  * sched_setparam() passes in -1 for its policy, to let the functions
  * it calls know not to change it.
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index e7cd7c148..ffb3e0cfc 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -848,7 +848,20 @@ static void update_curr(struct cfs_rq *cfs_rq)
 	curr->sum_exec_runtime += delta_exec;
 	schedstat_add(cfs_rq->exec_clock, delta_exec);
 
-	curr->vruntime += calc_delta_fair(delta_exec, curr);
+	unsigned long flags;
+	struct inactive_tgid *entry;
+	struct task_struct *task = container_of(curr, struct task_struct, se);
+	u64 penalty = 0;
+
+	spin_lock_irqsave(&inactive_tgid_lock, flags);
+	list_for_each_entry(entry, &inactive_tgid_list, list) {
+		if (entry->tgid == task->tgid) {
+			penalty = entry->penalty;
+			break;
+		}
+	}
+	spin_unlock_irqrestore(&inactive_tgid_lock, flags);
+	curr->vruntime += calc_delta_fair(delta_exec, curr) + penalty;
 	update_min_vruntime(cfs_rq);
 
 	if (entity_is_task(curr)) {
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 519a9343a..1ce2956b3 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -88,12 +88,21 @@
 struct rq;
 struct cpuidle_state;
 
+struct inactive_tgid {
+	pid_t tgid;
+	u64 penalty;
+	struct list_head list;
+};
+
 /* task_struct::on_rq states: */
 #define TASK_ON_RQ_QUEUED	1
 #define TASK_ON_RQ_MIGRATING	2
 
 extern __read_mostly int scheduler_running;
 
+extern struct list_head inactive_tgid_list;
+extern spinlock_t inactive_tgid_lock;
+
 extern unsigned long calc_load_update;
 extern atomic_long_t calc_load_tasks;
 
diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c
index 8671c37f0..e7de63c53 100644
--- a/kernel/sys_ni.c
+++ b/kernel/sys_ni.c
@@ -36,6 +36,7 @@ asmlinkage long sys_ni_syscall(void)
  * system calls.
  */
 
+COND_SYSCALL(set_process_inactive);
 COND_SYSCALL(io_setup);
 COND_SYSCALL_COMPAT(io_setup);
 COND_SYSCALL(io_destroy);