Merge tag 'sched-fifo-2020-08-04' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull sched/fifo updates from Ingo Molnar:
"This adds the sched_set_fifo*() encapsulation APIs to remove static
priority level knowledge from non-scheduler code.

The three APIs for non-scheduler code to set SCHED_FIFO are:

- sched_set_fifo()
- sched_set_fifo_low()
- sched_set_normal()

These are two FIFO priority levels: default (high), and a 'low'
priority level, plus sched_set_normal() to set the policy back to
non-SCHED_FIFO.

Since the changes affect a lot of non-scheduler code, we kept this in
a separate tree"

* tag 'sched-fifo-2020-08-04' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (24 commits)
sched,tracing: Convert to sched_set_fifo()
sched: Remove sched_set_*() return value
sched: Remove sched_setscheduler*() EXPORTs
sched,psi: Convert to sched_set_fifo_low()
sched,rcutorture: Convert to sched_set_fifo_low()
sched,rcuperf: Convert to sched_set_fifo_low()
sched,locktorture: Convert to sched_set_fifo()
sched,irq: Convert to sched_set_fifo()
sched,watchdog: Convert to sched_set_fifo()
sched,serial: Convert to sched_set_fifo()
sched,powerclamp: Convert to sched_set_fifo()
sched,ion: Convert to sched_set_normal()
sched,powercap: Convert to sched_set_fifo*()
sched,spi: Convert to sched_set_fifo*()
sched,mmc: Convert to sched_set_fifo*()
sched,ivtv: Convert to sched_set_fifo*()
sched,drm/scheduler: Convert to sched_set_fifo*()
sched,msm: Convert to sched_set_fifo*()
sched,psci: Convert to sched_set_fifo*()
sched,drbd: Convert to sched_set_fifo*()
...

arch/arm/common/bL_switcher.c

···
 static int bL_switcher_thread(void *arg)
 {
 	struct bL_thread *t = arg;
-	struct sched_param param = { .sched_priority = 1 };
 	int cluster;
 	bL_switch_completion_handler completer;
 	void *completer_cookie;
 
-	sched_setscheduler_nocheck(current, SCHED_FIFO, &param);
+	sched_set_fifo_low(current);
 	complete(&t->started);
 
 	do {

crypto/crypto_engine.c

···
 						       int (*cbk_do_batch)(struct crypto_engine *engine),
 						       bool rt, int qlen)
 {
-	struct sched_param param = { .sched_priority = MAX_RT_PRIO / 2 };
 	struct crypto_engine *engine;
 
 	if (!dev)
···
 
 	if (engine->rt) {
 		dev_info(dev, "will run requests pump with realtime priority\n");
-		sched_setscheduler(engine->kworker->task, SCHED_FIFO, &param);
+		sched_set_fifo(engine->kworker->task);
 	}
 
 	return engine;

drivers/acpi/acpi_pad.c

···
 static unsigned int round_robin_time = 1; /* second */
 static int power_saving_thread(void *data)
 {
-	struct sched_param param = {.sched_priority = 1};
 	int do_sleep;
 	unsigned int tsk_index = (unsigned long)data;
 	u64 last_jiffies = 0;
 
-	sched_setscheduler(current, SCHED_RR, &param);
+	sched_set_fifo_low(current);
 
 	while (!kthread_should_stop()) {
 		unsigned long expire_time;

drivers/block/drbd/drbd_receiver.c

···
 	unsigned int header_size = drbd_header_size(connection);
 	int expect   = header_size;
 	bool ping_timeout_active = false;
-	struct sched_param param = { .sched_priority = 2 };
 
-	rv = sched_setscheduler(current, SCHED_RR, &param);
-	if (rv < 0)
-		drbd_err(connection, "drbd_ack_receiver: ERROR set priority, ret=%d\n", rv);
+	sched_set_fifo_low(current);
 
 	while (get_t_state(thi) == RUNNING) {
 		drbd_thread_current_set_cpu(thi);

drivers/firmware/psci/psci_checker.c

···
 {
 	int cpu = (long)arg;
 	int i, nb_suspend = 0, nb_shallow_sleep = 0, nb_err = 0;
-	struct sched_param sched_priority = { .sched_priority = MAX_RT_PRIO-1 };
 	struct cpuidle_device *dev;
 	struct cpuidle_driver *drv;
 	/* No need for an actual callback, we just want to wake up the CPU. */
···
 	wait_for_completion(&suspend_threads_started);
 
 	/* Set maximum priority to preempt all other threads on this CPU. */
-	if (sched_setscheduler_nocheck(current, SCHED_FIFO, &sched_priority))
-		pr_warn("Failed to set suspend thread scheduler on CPU %d\n",
-			cpu);
+	sched_set_fifo(current);
 
 	dev = this_cpu_read(cpuidle_devices);
 	drv = cpuidle_get_cpu_driver(dev);
···
 	if (atomic_dec_return_relaxed(&nb_active_threads) == 0)
 		complete(&suspend_threads_done);
 
-	/* Give up on RT scheduling and wait for termination. */
-	sched_priority.sched_priority = 0;
-	if (sched_setscheduler_nocheck(current, SCHED_NORMAL, &sched_priority))
-		pr_warn("Failed to set suspend thread scheduler on CPU %d\n",
-			cpu);
 	for (;;) {
 		/* Needs to be set first to avoid missing a wakeup. */
 		set_current_state(TASK_INTERRUPTIBLE);

drivers/gpu/drm/drm_vblank_work.c

···
 
 int drm_vblank_worker_init(struct drm_vblank_crtc *vblank)
 {
-	struct sched_param param = {
-		.sched_priority = MAX_RT_PRIO - 1,
-	};
 	struct kthread_worker *worker;
 
 	INIT_LIST_HEAD(&vblank->pending_work);
···
 
 	vblank->worker = worker;
 
-	return sched_setscheduler(vblank->worker->task, SCHED_FIFO, &param);
+	sched_set_fifo(worker->task);
+	return 0;
 }

drivers/gpu/drm/msm/msm_drv.c

···
 	struct msm_kms *kms;
 	struct msm_mdss *mdss;
 	int ret, i;
-	struct sched_param param;
 
 	ddev = drm_dev_alloc(drv, dev);
 	if (IS_ERR(ddev)) {
···
 	ddev->mode_config.funcs = &mode_config_funcs;
 	ddev->mode_config.helper_private = &mode_config_helper_funcs;
 
-	/**
-	 * this priority was found during empiric testing to have appropriate
-	 * realtime scheduling to process display updates and interact with
-	 * other real time and normal priority task
-	 */
-	param.sched_priority = 16;
 	for (i = 0; i < priv->num_crtcs; i++) {
 		/* initialize event thread */
 		priv->event_thread[i].crtc_id = priv->crtcs[i]->base.id;
···
 			goto err_msm_uninit;
 		}
 
-		ret = sched_setscheduler(priv->event_thread[i].worker->task,
-					 SCHED_FIFO, &param);
-		if (ret)
-			dev_warn(dev, "event_thread set priority failed:%d\n",
-				 ret);
+		sched_set_fifo(priv->event_thread[i].worker->task);
 	}
 
 	ret = drm_vblank_init(ddev, priv->num_crtcs);

drivers/gpu/drm/scheduler/sched_main.c

···
  */
 static int drm_sched_main(void *param)
 {
-	struct sched_param sparam = {.sched_priority = 1};
 	struct drm_gpu_scheduler *sched = (struct drm_gpu_scheduler *)param;
 	int r;
 
-	sched_setscheduler(current, SCHED_FIFO, &sparam);
+	sched_set_fifo_low(current);
 
 	while (!kthread_should_stop()) {
 		struct drm_sched_entity *entity = NULL;

drivers/media/pci/ivtv/ivtv-driver.c

···
  */
 static int ivtv_init_struct1(struct ivtv *itv)
 {
-	struct sched_param param = { .sched_priority = 99 };
-
 	itv->base_addr = pci_resource_start(itv->pdev, 0);
 	itv->enc_mbox.max_mbox = 2; /* the encoder has 3 mailboxes (0-2) */
 	itv->dec_mbox.max_mbox = 1; /* the decoder has 2 mailboxes (0-1) */
···
 		return -1;
 	}
 	/* must use the FIFO scheduler as it is realtime sensitive */
-	sched_setscheduler(itv->irq_worker_task, SCHED_FIFO, &param);
+	sched_set_fifo(itv->irq_worker_task);
 
 	kthread_init_work(&itv->irq_work, ivtv_irq_work_handler);
 

drivers/mmc/core/sdio_irq.c

···
 static int sdio_irq_thread(void *_host)
 {
 	struct mmc_host *host = _host;
-	struct sched_param param = { .sched_priority = 1 };
 	unsigned long period, idle_period;
 	int ret;
 
-	sched_setscheduler(current, SCHED_FIFO, &param);
+	sched_set_fifo_low(current);
 
 	/*
 	 * We want to allow for SDIO cards to work even on non SDIO

drivers/platform/chrome/cros_ec_spi.c

···
 static int cros_ec_spi_devm_high_pri_alloc(struct device *dev,
 					   struct cros_ec_spi *ec_spi)
 {
-	struct sched_param sched_priority = {
-		.sched_priority = MAX_RT_PRIO / 2,
-	};
 	int err;
 
 	ec_spi->high_pri_worker =
···
 	if (err)
 		return err;
 
-	err = sched_setscheduler_nocheck(ec_spi->high_pri_worker->task,
-					 SCHED_FIFO, &sched_priority);
-	if (err)
-		dev_err(dev, "Can't set cros_ec high pri priority: %d\n", err);
-	return err;
+	sched_set_fifo(ec_spi->high_pri_worker->task);
+
+	return 0;
 }
 
 static int cros_ec_spi_probe(struct spi_device *spi)

drivers/powercap/idle_inject.c

···
  */
 static void idle_inject_setup(unsigned int cpu)
 {
-	struct sched_param param = { .sched_priority = MAX_USER_RT_PRIO / 2 };
-
-	sched_setscheduler(current, SCHED_FIFO, &param);
+	sched_set_fifo(current);
 }
 
 /**

drivers/spi/spi.c

···
  */
 static void spi_set_thread_rt(struct spi_controller *ctlr)
 {
-	struct sched_param param = { .sched_priority = MAX_RT_PRIO / 2 };
-
 	dev_info(&ctlr->dev,
 		"will run message pump with realtime priority\n");
-	sched_setscheduler(ctlr->kworker->task, SCHED_FIFO, &param);
+	sched_set_fifo(ctlr->kworker->task);
 }
 
 static int spi_init_queue(struct spi_controller *ctlr)

drivers/staging/android/ion/ion_heap.c

···
 
 int ion_heap_init_deferred_free(struct ion_heap *heap)
 {
-	struct sched_param param = { .sched_priority = 0 };
-
 	INIT_LIST_HEAD(&heap->free_list);
 	init_waitqueue_head(&heap->waitqueue);
 	heap->task = kthread_run(ion_heap_deferred_free, heap,
···
 		       __func__);
 		return PTR_ERR_OR_ZERO(heap->task);
 	}
-	sched_setscheduler(heap->task, SCHED_IDLE, &param);
+	sched_set_normal(heap->task, 19);
 
 	return 0;
 }

drivers/thermal/intel/intel_powerclamp.c

···
 				  */
 static bool clamping;
 
-static const struct sched_param sparam = {
-	.sched_priority = MAX_USER_RT_PRIO / 2,
-};
 struct powerclamp_worker_data {
 	struct kthread_worker *worker;
 	struct kthread_work balancing_work;
···
 	w_data->cpu = cpu;
 	w_data->clamping = true;
 	set_bit(cpu, cpu_clamping_mask);
-	sched_setscheduler(worker->task, SCHED_FIFO, &sparam);
+	sched_set_fifo(worker->task);
 	kthread_init_work(&w_data->balancing_work, clamp_balancing_func);
 	kthread_init_delayed_work(&w_data->idle_injection_work,
 				  clamp_idle_injection_func);

drivers/tty/serial/sc16is7xx.c

···
 			   const struct sc16is7xx_devtype *devtype,
 			   struct regmap *regmap, int irq)
 {
-	struct sched_param sched_param = { .sched_priority = MAX_RT_PRIO / 2 };
 	unsigned long freq = 0, *pfreq = dev_get_platdata(dev);
 	unsigned int val;
 	u32 uartclk = 0;
···
 		ret = PTR_ERR(s->kworker_task);
 		goto out_clk;
 	}
-	sched_setscheduler(s->kworker_task, SCHED_FIFO, &sched_param);
+	sched_set_fifo(s->kworker_task);
 
 #ifdef CONFIG_GPIOLIB
 	if (devtype->nr_gpio) {

drivers/watchdog/watchdog_dev.c

···
 int __init watchdog_dev_init(void)
 {
 	int err;
-	struct sched_param param = {.sched_priority = MAX_RT_PRIO - 1,};
 
 	watchdog_kworker = kthread_create_worker(0, "watchdogd");
 	if (IS_ERR(watchdog_kworker)) {
 		pr_err("Failed to create watchdog kworker\n");
 		return PTR_ERR(watchdog_kworker);
 	}
-	sched_setscheduler(watchdog_kworker->task, SCHED_FIFO, &param);
+	sched_set_fifo(watchdog_kworker->task);
 
 	err = class_register(&watchdog_class);
 	if (err < 0) {

include/linux/sched.h

···
 extern int available_idle_cpu(int cpu);
 extern int sched_setscheduler(struct task_struct *, int, const struct sched_param *);
 extern int sched_setscheduler_nocheck(struct task_struct *, int, const struct sched_param *);
+extern void sched_set_fifo(struct task_struct *p);
+extern void sched_set_fifo_low(struct task_struct *p);
+extern void sched_set_normal(struct task_struct *p, int nice);
 extern int sched_setattr(struct task_struct *, const struct sched_attr *);
 extern int sched_setattr_nocheck(struct task_struct *, const struct sched_attr *);
 extern struct task_struct *idle_task(int cpu);

kernel/irq/manage.c

···
 setup_irq_thread(struct irqaction *new, unsigned int irq, bool secondary)
 {
 	struct task_struct *t;
-	struct sched_param param = {
-		.sched_priority = MAX_USER_RT_PRIO/2,
-	};
 
 	if (!secondary) {
 		t = kthread_create(irq_thread, new, "irq/%d-%s", irq,
···
 	} else {
 		t = kthread_create(irq_thread, new, "irq/%d-s-%s", irq,
 				   new->name);
-		param.sched_priority -= 1;
 	}
 
 	if (IS_ERR(t))
 		return PTR_ERR(t);
 
-	sched_setscheduler_nocheck(t, SCHED_FIFO, &param);
+	sched_set_fifo(t);
 
 	/*
 	 * We keep the reference to the task struct even if

kernel/locking/locktorture.c

···
 
 static void torture_rtmutex_boost(struct torture_random_state *trsp)
 {
-	int policy;
-	struct sched_param param;
 	const unsigned int factor = 50000; /* yes, quite arbitrary */
 
 	if (!rt_task(current)) {
···
 		 */
 		if (trsp && !(torture_random(trsp) %
 			      (cxt.nrealwriters_stress * factor))) {
-			policy = SCHED_FIFO;
-			param.sched_priority = MAX_RT_PRIO - 1;
+			sched_set_fifo(current);
 		} else /* common case, do nothing */
 			return;
 	} else {
···
 		 */
 		if (!trsp || !(torture_random(trsp) %
 			       (cxt.nrealwriters_stress * factor * 2))) {
-			policy = SCHED_NORMAL;
-			param.sched_priority = 0;
+			sched_set_normal(current, 0);
 		} else /* common case, do nothing */
 			return;
 	}
-
-	sched_setscheduler_nocheck(current, policy, &param);
 }
 
 static void torture_rtmutex_delay(struct torture_random_state *trsp)

kernel/rcu/rcuperf.c

···
 	int i_max;
 	long me = (long)arg;
 	struct rcu_head *rhp = NULL;
-	struct sched_param sp;
 	bool started = false, done = false, alldone = false;
 	u64 t;
 	u64 *wdp;
···
 	VERBOSE_PERFOUT_STRING("rcu_perf_writer task started");
 	WARN_ON(!wdpp);
 	set_cpus_allowed_ptr(current, cpumask_of(me % nr_cpu_ids));
-	sp.sched_priority = 1;
-	sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
+	sched_set_fifo_low(current);
 
 	if (holdoff)
 		schedule_timeout_uninterruptible(holdoff * HZ);
···
 			started = true;
 		if (!done && i >= MIN_MEAS) {
 			done = true;
-			sp.sched_priority = 0;
-			sched_setscheduler_nocheck(current,
-						   SCHED_NORMAL, &sp);
+			sched_set_normal(current, 0);
 			pr_alert("%s%s rcu_perf_writer %ld has %d measurements\n",
 				 perf_type, PERF_FLAG, me, MIN_MEAS);
 			if (atomic_inc_return(&n_rcu_perf_writer_finished) >=

kernel/rcu/rcutorture.c

···
 	unsigned long endtime;
 	unsigned long oldstarttime;
 	struct rcu_boost_inflight rbi = { .inflight = 0 };
-	struct sched_param sp;
 
 	VERBOSE_TOROUT_STRING("rcu_torture_boost started");
 
 	/* Set real-time priority. */
-	sp.sched_priority = 1;
-	if (sched_setscheduler(current, SCHED_FIFO, &sp) < 0) {
-		VERBOSE_TOROUT_STRING("rcu_torture_boost RT prio failed!");
-		n_rcu_torture_boost_rterror++;
-	}
+	sched_set_fifo_low(current);
 
 	init_rcu_head_on_stack(&rbi.rcu);
 	/* Each pass through the following loop does one boost-test cycle. */

kernel/sched/core.c

···
  * @policy: new policy.
  * @param: structure containing the new RT priority.
  *
+ * Use sched_set_fifo(), read its comment.
+ *
  * Return: 0 on success. An error code otherwise.
  *
  * NOTE that the task may be already dead.
···
 {
 	return _sched_setscheduler(p, policy, param, true);
 }
-EXPORT_SYMBOL_GPL(sched_setscheduler);
 
 int sched_setattr(struct task_struct *p, const struct sched_attr *attr)
 {
 	return __sched_setscheduler(p, attr, true, true);
 }
-EXPORT_SYMBOL_GPL(sched_setattr);
 
 int sched_setattr_nocheck(struct task_struct *p, const struct sched_attr *attr)
 {
···
 {
 	return _sched_setscheduler(p, policy, param, false);
 }
-EXPORT_SYMBOL_GPL(sched_setscheduler_nocheck);
+
+/*
+ * SCHED_FIFO is a broken scheduler model; that is, it is fundamentally
+ * incapable of resource management, which is the one thing an OS really should
+ * be doing.
+ *
+ * This is of course the reason it is limited to privileged users only.
+ *
+ * Worse still; it is fundamentally impossible to compose static priority
+ * workloads. You cannot take two correctly working static prio workloads
+ * and smash them together and still expect them to work.
+ *
+ * For this reason 'all' FIFO tasks the kernel creates are basically at:
+ *
+ *   MAX_RT_PRIO / 2
+ *
+ * The administrator _MUST_ configure the system, the kernel simply doesn't
+ * know enough information to make a sensible choice.
+ */
+void sched_set_fifo(struct task_struct *p)
+{
+	struct sched_param sp = { .sched_priority = MAX_RT_PRIO / 2 };
+	WARN_ON_ONCE(sched_setscheduler_nocheck(p, SCHED_FIFO, &sp) != 0);
+}
+EXPORT_SYMBOL_GPL(sched_set_fifo);
+
+/*
+ * For when you don't much care about FIFO, but want to be above SCHED_NORMAL.
+ */
+void sched_set_fifo_low(struct task_struct *p)
+{
+	struct sched_param sp = { .sched_priority = 1 };
+	WARN_ON_ONCE(sched_setscheduler_nocheck(p, SCHED_FIFO, &sp) != 0);
+}
+EXPORT_SYMBOL_GPL(sched_set_fifo_low);
+
+void sched_set_normal(struct task_struct *p, int nice)
+{
+	struct sched_attr attr = {
+		.sched_policy = SCHED_NORMAL,
+		.sched_nice = nice,
+	};
+	WARN_ON_ONCE(sched_setattr_nocheck(p, &attr) != 0);
+}
+EXPORT_SYMBOL_GPL(sched_set_normal);
 
 static int
 do_sched_setscheduler(pid_t pid, int policy, struct sched_param __user *param)

kernel/sched/psi.c

···
 static int psi_poll_worker(void *data)
 {
 	struct psi_group *group = (struct psi_group *)data;
-	struct sched_param param = {
-		.sched_priority = 1,
-	};
 
-	sched_setscheduler_nocheck(current, SCHED_FIFO, &param);
+	sched_set_fifo_low(current);
 
 	while (true) {
 		wait_event_interruptible(group->poll_wait,

kernel/trace/ring_buffer_benchmark.c

···
 static int producer_nice = MAX_NICE;
 static int consumer_nice = MAX_NICE;
 
-static int producer_fifo = -1;
-static int consumer_fifo = -1;
+static int producer_fifo;
+static int consumer_fifo;
 
 module_param(producer_nice, int, 0644);
 MODULE_PARM_DESC(producer_nice, "nice prio for producer");
···
 MODULE_PARM_DESC(consumer_nice, "nice prio for consumer");
 
 module_param(producer_fifo, int, 0644);
-MODULE_PARM_DESC(producer_fifo, "fifo prio for producer");
+MODULE_PARM_DESC(producer_fifo, "use fifo for producer: 0 - disabled, 1 - low prio, 2 - fifo");
 
 module_param(consumer_fifo, int, 0644);
-MODULE_PARM_DESC(consumer_fifo, "fifo prio for consumer");
+MODULE_PARM_DESC(consumer_fifo, "use fifo for consumer: 0 - disabled, 1 - low prio, 2 - fifo");
 
 static int read_events;
 
···
 		trace_printk("ERROR!\n");
 
 	if (!disable_reader) {
-		if (consumer_fifo < 0)
+		if (consumer_fifo)
+			trace_printk("Running Consumer at SCHED_FIFO %s\n",
+				     consumer_fifo == 1 ? "low" : "high");
+		else
 			trace_printk("Running Consumer at nice: %d\n",
 				     consumer_nice);
-		else
-			trace_printk("Running Consumer at SCHED_FIFO %d\n",
-				     consumer_fifo);
 	}
-	if (producer_fifo < 0)
+	if (producer_fifo)
+		trace_printk("Running Producer at SCHED_FIFO %s\n",
+			     producer_fifo == 1 ? "low" : "high");
+	else
 		trace_printk("Running Producer at nice: %d\n",
 			     producer_nice);
-	else
-		trace_printk("Running Producer at SCHED_FIFO %d\n",
-			     producer_fifo);
 
 	/* Let the user know that the test is running at low priority */
-	if (producer_fifo < 0 && consumer_fifo < 0 &&
+	if (!producer_fifo && !consumer_fifo &&
 	    producer_nice == MAX_NICE && consumer_nice == MAX_NICE)
 		trace_printk("WARNING!!! This test is running at lowest priority.\n");
 
···
 	 * Run them as low-prio background tasks by default:
 	 */
 	if (!disable_reader) {
-		if (consumer_fifo >= 0) {
-			struct sched_param param = {
-				.sched_priority = consumer_fifo
-			};
-			sched_setscheduler(consumer, SCHED_FIFO, &param);
-		} else
+		if (consumer_fifo >= 2)
+			sched_set_fifo(consumer);
+		else if (consumer_fifo == 1)
+			sched_set_fifo_low(consumer);
+		else
 			set_user_nice(consumer, consumer_nice);
 	}
 
-	if (producer_fifo >= 0) {
-		struct sched_param param = {
-			.sched_priority = producer_fifo
-		};
-		sched_setscheduler(producer, SCHED_FIFO, &param);
-	} else
+	if (producer_fifo >= 2)
+		sched_set_fifo(producer);
+	else if (producer_fifo == 1)
+		sched_set_fifo_low(producer);
+	else
 		set_user_nice(producer, producer_nice);
 
 	return 0;