Revision 63cae12bce9861cec309798d34701cf3da20bc71 authored by Peter Zijlstra on 09 December 2016, 13:59:00 UTC, committed by Ingo Molnar on 14 January 2017, 09:56:10 UTC
There is problem with installing an event in a task that is 'stuck' on
an offline CPU.

Blocked tasks are not dis-assosciated from offlined CPUs, after all, a
blocked task doesn't run and doesn't require a CPU etc.. Only on
wakeup do we ammend the situation and place the task on a available
CPU.

If we hit such a task with perf_install_in_context() we'll loop until
either that task wakes up or the CPU comes back online, if the task
waking depends on the event being installed, we're stuck.

While looking into this issue, I also spotted another problem, if we
hit a task with perf_install_in_context() that is in the middle of
being migrated, that is we observe the old CPU before sending the IPI,
but run the IPI (on the old CPU) while the task is already running on
the new CPU, things also go sideways.

Rework things to rely on task_curr() -- outside of rq->lock -- which
is rather tricky. Imagine the following scenario where we're trying to
install the first event into our task 't':

CPU0            CPU1            CPU2

                (current == t)

t->perf_event_ctxp[] = ctx;
smp_mb();
cpu = task_cpu(t);

                switch(t, n);
                                migrate(t, 2);
                                switch(p, t);

                                ctx = t->perf_event_ctxp[]; // must not be NULL

smp_function_call(cpu, ..);

                generic_exec_single()
                  func();
                    spin_lock(ctx->lock);
                    if (task_curr(t)) // false

                    add_event_to_ctx();
                    spin_unlock(ctx->lock);

                                perf_event_context_sched_in();
                                  spin_lock(ctx->lock);
                                  // sees event

So its CPU0's store of t->perf_event_ctxp[] that must not go 'missing'.
Because if CPU2's load of that variable were to observe NULL, it would
not try to schedule the ctx and we'd have a task running without its
counter, which would be 'bad'.

As long as we observe !NULL, we'll acquire ctx->lock. If we acquire it
first and not see the event yet, then CPU0 must observe task_curr()
and retry. If the install happens first, then we must see the event on
sched-in and all is well.

I think we can translate the first part (until the 'must not be NULL')
of the scenario to a litmus test like:

  C C-peterz

  {
  }

  P0(int *x, int *y)
  {
          int r1;

          WRITE_ONCE(*x, 1);
          smp_mb();
          r1 = READ_ONCE(*y);
  }

  P1(int *y, int *z)
  {
          WRITE_ONCE(*y, 1);
          smp_store_release(z, 1);
  }

  P2(int *x, int *z)
  {
          int r1;
          int r2;

          r1 = smp_load_acquire(z);
	  smp_mb();
          r2 = READ_ONCE(*x);
  }

  exists
  (0:r1=0 /\ 2:r1=1 /\ 2:r2=0)

Where:
  x is perf_event_ctxp[],
  y is our tasks's CPU, and
  z is our task being placed on the rq of CPU2.

The P0 smp_mb() is the one added by this patch, ordering the store to
perf_event_ctxp[] from find_get_context() and the load of task_cpu()
in task_function_call().

The smp_store_release/smp_load_acquire model the RCpc locking of the
rq->lock and the smp_mb() of P2 is the context switch switching from
whatever CPU2 was running to our task 't'.

This litmus test evaluates into:

  Test C-peterz Allowed
  States 7
  0:r1=0; 2:r1=0; 2:r2=0;
  0:r1=0; 2:r1=0; 2:r2=1;
  0:r1=0; 2:r1=1; 2:r2=1;
  0:r1=1; 2:r1=0; 2:r2=0;
  0:r1=1; 2:r1=0; 2:r2=1;
  0:r1=1; 2:r1=1; 2:r2=0;
  0:r1=1; 2:r1=1; 2:r2=1;
  No
  Witnesses
  Positive: 0 Negative: 7
  Condition exists (0:r1=0 /\ 2:r1=1 /\ 2:r2=0)
  Observation C-peterz Never 0 7
  Hash=e427f41d9146b2a5445101d3e2fcaa34

And the strong and weak model agree.

Reported-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: Will Deacon <will.deacon@arm.com>
Cc: jeremy.linton@arm.com
Link: http://lkml.kernel.org/r/20161209135900.GU3174@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
1 parent ad5013d
Raw File
leds-gpio.c
/*
 * LEDs driver for GPIOs
 *
 * Copyright (C) 2007 8D Technologies inc.
 * Raphael Assenat <raph@8d.com>
 * Copyright (C) 2008 Freescale Semiconductor, Inc.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 */
#include <linux/err.h>
#include <linux/gpio.h>
#include <linux/gpio/consumer.h>
#include <linux/kernel.h>
#include <linux/leds.h>
#include <linux/module.h>
#include <linux/of.h>
#include <linux/platform_device.h>
#include <linux/property.h>
#include <linux/slab.h>

struct gpio_led_data {
	struct led_classdev cdev;
	struct gpio_desc *gpiod;
	u8 can_sleep;
	u8 blinking;
	gpio_blink_set_t platform_gpio_blink_set;
};

static inline struct gpio_led_data *
			cdev_to_gpio_led_data(struct led_classdev *led_cdev)
{
	return container_of(led_cdev, struct gpio_led_data, cdev);
}

static void gpio_led_set(struct led_classdev *led_cdev,
	enum led_brightness value)
{
	struct gpio_led_data *led_dat = cdev_to_gpio_led_data(led_cdev);
	int level;

	if (value == LED_OFF)
		level = 0;
	else
		level = 1;

	if (led_dat->blinking) {
		led_dat->platform_gpio_blink_set(led_dat->gpiod, level,
						 NULL, NULL);
		led_dat->blinking = 0;
	} else {
		if (led_dat->can_sleep)
			gpiod_set_value_cansleep(led_dat->gpiod, level);
		else
			gpiod_set_value(led_dat->gpiod, level);
	}
}

static int gpio_led_set_blocking(struct led_classdev *led_cdev,
	enum led_brightness value)
{
	gpio_led_set(led_cdev, value);
	return 0;
}

static int gpio_blink_set(struct led_classdev *led_cdev,
	unsigned long *delay_on, unsigned long *delay_off)
{
	struct gpio_led_data *led_dat = cdev_to_gpio_led_data(led_cdev);

	led_dat->blinking = 1;
	return led_dat->platform_gpio_blink_set(led_dat->gpiod, GPIO_LED_BLINK,
						delay_on, delay_off);
}

static int create_gpio_led(const struct gpio_led *template,
	struct gpio_led_data *led_dat, struct device *parent,
	gpio_blink_set_t blink_set)
{
	int ret, state;

	led_dat->gpiod = template->gpiod;
	if (!led_dat->gpiod) {
		/*
		 * This is the legacy code path for platform code that
		 * still uses GPIO numbers. Ultimately we would like to get
		 * rid of this block completely.
		 */
		unsigned long flags = GPIOF_OUT_INIT_LOW;

		/* skip leds that aren't available */
		if (!gpio_is_valid(template->gpio)) {
			dev_info(parent, "Skipping unavailable LED gpio %d (%s)\n",
					template->gpio, template->name);
			return 0;
		}

		if (template->active_low)
			flags |= GPIOF_ACTIVE_LOW;

		ret = devm_gpio_request_one(parent, template->gpio, flags,
					    template->name);
		if (ret < 0)
			return ret;

		led_dat->gpiod = gpio_to_desc(template->gpio);
		if (!led_dat->gpiod)
			return -EINVAL;
	}

	led_dat->cdev.name = template->name;
	led_dat->cdev.default_trigger = template->default_trigger;
	led_dat->can_sleep = gpiod_cansleep(led_dat->gpiod);
	if (!led_dat->can_sleep)
		led_dat->cdev.brightness_set = gpio_led_set;
	else
		led_dat->cdev.brightness_set_blocking = gpio_led_set_blocking;
	led_dat->blinking = 0;
	if (blink_set) {
		led_dat->platform_gpio_blink_set = blink_set;
		led_dat->cdev.blink_set = gpio_blink_set;
	}
	if (template->default_state == LEDS_GPIO_DEFSTATE_KEEP) {
		state = gpiod_get_value_cansleep(led_dat->gpiod);
		if (state < 0)
			return state;
	} else {
		state = (template->default_state == LEDS_GPIO_DEFSTATE_ON);
	}
	led_dat->cdev.brightness = state ? LED_FULL : LED_OFF;
	if (!template->retain_state_suspended)
		led_dat->cdev.flags |= LED_CORE_SUSPENDRESUME;
	if (template->panic_indicator)
		led_dat->cdev.flags |= LED_PANIC_INDICATOR;

	ret = gpiod_direction_output(led_dat->gpiod, state);
	if (ret < 0)
		return ret;

	return devm_led_classdev_register(parent, &led_dat->cdev);
}

struct gpio_leds_priv {
	int num_leds;
	struct gpio_led_data leds[];
};

static inline int sizeof_gpio_leds_priv(int num_leds)
{
	return sizeof(struct gpio_leds_priv) +
		(sizeof(struct gpio_led_data) * num_leds);
}

static struct gpio_leds_priv *gpio_leds_create(struct platform_device *pdev)
{
	struct device *dev = &pdev->dev;
	struct fwnode_handle *child;
	struct gpio_leds_priv *priv;
	int count, ret;

	count = device_get_child_node_count(dev);
	if (!count)
		return ERR_PTR(-ENODEV);

	priv = devm_kzalloc(dev, sizeof_gpio_leds_priv(count), GFP_KERNEL);
	if (!priv)
		return ERR_PTR(-ENOMEM);

	device_for_each_child_node(dev, child) {
		struct gpio_led_data *led_dat = &priv->leds[priv->num_leds];
		struct gpio_led led = {};
		const char *state = NULL;
		struct device_node *np = to_of_node(child);

		led.gpiod = devm_get_gpiod_from_child(dev, NULL, child);
		if (IS_ERR(led.gpiod)) {
			fwnode_handle_put(child);
			return ERR_CAST(led.gpiod);
		}

		ret = fwnode_property_read_string(child, "label", &led.name);
		if (ret && IS_ENABLED(CONFIG_OF) && np)
			led.name = np->name;
		if (!led.name) {
			fwnode_handle_put(child);
			return ERR_PTR(-EINVAL);
		}

		fwnode_property_read_string(child, "linux,default-trigger",
					    &led.default_trigger);

		if (!fwnode_property_read_string(child, "default-state",
						 &state)) {
			if (!strcmp(state, "keep"))
				led.default_state = LEDS_GPIO_DEFSTATE_KEEP;
			else if (!strcmp(state, "on"))
				led.default_state = LEDS_GPIO_DEFSTATE_ON;
			else
				led.default_state = LEDS_GPIO_DEFSTATE_OFF;
		}

		if (fwnode_property_present(child, "retain-state-suspended"))
			led.retain_state_suspended = 1;
		if (fwnode_property_present(child, "panic-indicator"))
			led.panic_indicator = 1;

		ret = create_gpio_led(&led, led_dat, dev, NULL);
		if (ret < 0) {
			fwnode_handle_put(child);
			return ERR_PTR(ret);
		}
		led_dat->cdev.dev->of_node = np;
		priv->num_leds++;
	}

	return priv;
}

static const struct of_device_id of_gpio_leds_match[] = {
	{ .compatible = "gpio-leds", },
	{},
};

MODULE_DEVICE_TABLE(of, of_gpio_leds_match);

static int gpio_led_probe(struct platform_device *pdev)
{
	struct gpio_led_platform_data *pdata = dev_get_platdata(&pdev->dev);
	struct gpio_leds_priv *priv;
	int i, ret = 0;

	if (pdata && pdata->num_leds) {
		priv = devm_kzalloc(&pdev->dev,
				sizeof_gpio_leds_priv(pdata->num_leds),
					GFP_KERNEL);
		if (!priv)
			return -ENOMEM;

		priv->num_leds = pdata->num_leds;
		for (i = 0; i < priv->num_leds; i++) {
			ret = create_gpio_led(&pdata->leds[i],
					      &priv->leds[i],
					      &pdev->dev, pdata->gpio_blink_set);
			if (ret < 0)
				return ret;
		}
	} else {
		priv = gpio_leds_create(pdev);
		if (IS_ERR(priv))
			return PTR_ERR(priv);
	}

	platform_set_drvdata(pdev, priv);

	return 0;
}

static void gpio_led_shutdown(struct platform_device *pdev)
{
	struct gpio_leds_priv *priv = platform_get_drvdata(pdev);
	int i;

	for (i = 0; i < priv->num_leds; i++) {
		struct gpio_led_data *led = &priv->leds[i];

		gpio_led_set(&led->cdev, LED_OFF);
	}
}

static struct platform_driver gpio_led_driver = {
	.probe		= gpio_led_probe,
	.shutdown	= gpio_led_shutdown,
	.driver		= {
		.name	= "leds-gpio",
		.of_match_table = of_gpio_leds_match,
	},
};

module_platform_driver(gpio_led_driver);

MODULE_AUTHOR("Raphael Assenat <raph@8d.com>, Trent Piepho <tpiepho@freescale.com>");
MODULE_DESCRIPTION("GPIO LED driver");
MODULE_LICENSE("GPL");
MODULE_ALIAS("platform:leds-gpio");
back to top