path: root/kernel/events/ring_buffer.c

// SPDX-License-Identifier: GPL-2.0
/*
 * Performance events ring-buffer code:
 *
 *  Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de>
 *  Copyright (C) 2008-2011 Red Hat, Inc., Ingo Molnar
 *  Copyright (C) 2008-2011 Red Hat, Inc., Peter Zijlstra
 *  Copyright  ©  2009 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
 */

#include <linux/perf_event.h>
#include <linux/vmalloc.h>
#include <linux/slab.h>
#include <linux/circ_buf.h>
#include <linux/poll.h>
#include <linux/nospec.h>

#include "internal.h"

static void perf_output_wakeup(struct perf_output_handle *handle)
{
	atomic_set(&handle->rb->poll, EPOLLIN | EPOLLRDNORM);

	handle->event->pending_wakeup = 1;

	if (*perf_event_fasync(handle->event) && !handle->event->pending_kill)
		handle->event->pending_kill = POLL_IN;

	irq_work_queue(&handle->event->pending_irq);
}

/*
 * We need to ensure a later event_id doesn't publish a head when a former
 * event isn't done writing. However since we need to deal with NMIs we
 * cannot fully serialize things.
 *
 * We only publish the head (and generate a wakeup) when the outer-most
 * event completes.
 */
static void perf_output_get_handle(struct perf_output_handle *handle)
{
	struct perf_buffer *rb = handle->rb;

	preempt_disable();

	/*
	 * Avoid an explicit LOAD/STORE such that architectures with memops
	 * can use them.
	 */
	(*(volatile unsigned int *)&rb->nest)++;
	handle->wakeup = local_read(&rb->wakeup);
}

static void perf_output_put_handle(struct perf_output_handle *handle)
{
	struct perf_buffer *rb = handle->rb;
	unsigned long head;
	unsigned int nest;

	/*
	 * If this isn't the outermost nesting, we don't have to update
	 * @rb->user_page->data_head.
	 */
	nest = READ_ONCE(rb->nest);
	if (nest > 1) {
		WRITE_ONCE(rb->nest, nest - 1);
		goto out;
	}

again:
	/*
	 * In order to avoid publishing a head value that goes backwards,
	 * we must ensure the load of @rb->head happens after we've
	 * incremented @rb->nest.
	 *
	 * Otherwise we can observe a @rb->head value before one published
	 * by an IRQ/NMI happening between the load and the increment.
	 */
	barrier();
	head = local_read(&rb->head);

	/*
	 * IRQ/NMI can happen here and advance @rb->head, causing our
	 * load above to be stale.
	 */

	/*
	 * Since the mmap() consumer (userspace) can run on a different CPU:
	 *
	 *   kernel				user
	 *
	 *   if (LOAD ->data_tail) {		LOAD ->data_head
	 *			(A)		smp_rmb()	(C)
	 *	STORE $data			LOAD $data
	 *	smp_wmb()	(B)		smp_mb()	(D)
	 *	STORE ->data_head		STORE ->data_tail
	 *   }
	 *
	 * Where A pairs with D, and B pairs with C.
	 *
	 * In our case (A) is a control dependency that separates the load of
	 * the ->data_tail and the stores of $data. In case ->data_tail
	 * indicates there is no room in the buffer to store $data we do not.
	 *
	 * D needs to be a full barrier since it separates the data READ
	 * from the tail WRITE.
	 *
	 * For B a WMB is sufficient since it separates two WRITEs, and for C
	 * an RMB is sufficient since it separates two READs.
	 *
	 * See perf_output_begin().
	 */
	smp_wmb(); /* B, matches C */
	WRITE_ONCE(rb->user_page->data_head, head);

	/*
	 * We must publish the head before decrementing the nest count,
	 * otherwise an IRQ/NMI can publish a more recent head value and our
	 * write will (temporarily) publish a stale value.
	 */
	barrier();
	WRITE_ONCE(rb->nest, 0);

	/*
	 * Ensure we decrement @rb->nest before we validate the @rb->head.
	 * Otherwise we cannot be sure we caught the 'last' nested update.
	 */
	barrier();
	if (unlikely(head != local_read(&rb->head))) {
		WRITE_ONCE(rb->nest, 1);
		goto again;
	}

	if (handle->wakeup != local_read(&rb->wakeup))
		perf_output_wakeup(handle);

out:
	preempt_enable();
}

static __always_inline bool
ring_buffer_has_space(unsigned long head, unsigned long tail,
		      unsigned long data_size, unsigned int size,
		      bool backward)
{
	if (!backward)
		return CIRC_SPACE(head, tail, data_size) >= size;
	else
		return CIRC_SPACE(tail, head, data_size) >= size;
}

static __always_inline int
__perf_output_begin(struct perf_output_handle *handle,
		    struct perf_sample_data *data,
		    struct perf_event *event, unsigned int size,
		    bool backward)
{
	struct perf_buffer *rb;
	unsigned long tail, offset, head;
	int have_lost, page_shift;
	struct {
		struct perf_event_header header;
		u64			 id;
		u64			 lost;
	} lost_event;

	rcu_read_lock();
	/*
	 * For inherited events we send all the output towards the parent.
	 */
	if (event->parent)
		event = event->parent;

	rb = rcu_dereference(event->rb);
	if (unlikely(!rb))
		goto out;

	if (unlikely(rb->paused)) {
		if (rb->nr_pages) {
			local_inc(&rb->lost);
			atomic64_inc(&event->lost_samples);
		}
		goto out;
	}

	handle->rb    = rb;
	handle->event = event;
	handle->flags = 0;

	</