path: root/drivers/infiniband/hw/hfi1/affinity.c

// SPDX-License-Identifier: GPL-2.0 OR BSD-3-Clause
/*
 * Copyright(c) 2015 - 2020 Intel Corporation.
 */

#include <linux/topology.h>
#include <linux/cpumask.h>
#include <linux/interrupt.h>
#include <linux/numa.h>

#include "hfi.h"
#include "affinity.h"
#include "sdma.h"
#include "trace.h"

struct hfi1_affinity_node_list node_affinity = {
	.list = LIST_HEAD_INIT(node_affinity.list),
	.lock = __MUTEX_INITIALIZER(node_affinity.lock)
};

/* Name of IRQ types, indexed by enum irq_type */
static const char * const irq_type_names[] = {
	"SDMA",
	"RCVCTXT",
	"NETDEVCTXT",
	"GENERAL",
	"OTHER",
};

/* Per NUMA node count of HFI devices */
static unsigned int *hfi1_per_node_cntr;

static inline void init_cpu_mask_set(struct cpu_mask_set *set)
{
	cpumask_clear(&set->mask);
	cpumask_clear(&set->used);
	set->gen = 0;
}

/* Increment generation of CPU set if needed */
static void _cpu_mask_set_gen_inc(struct cpu_mask_set *set)
{
	if (cpumask_equal(&set->mask, &set->used)) {
		/*
		 * We've used up all the CPUs, bump up the generation
		 * and reset the 'used' map
		 */
		set->gen++;
		cpumask_clear(&set->used);
	}
}

static void _cpu_mask_set_gen_dec(struct cpu_mask_set *set)
{
	if (cpumask_empty(&set->used) && set->gen) {
		set->gen--;
		cpumask_copy(&set->used, &set->mask);
	}
}

/* Get the first CPU from the list of unused CPUs in a CPU set data structure */
static int cpu_mask_set_get_first(struct cpu_mask_set *set, cpumask_var_t diff)
{
	int cpu;

	if (!diff || !set)
		return -EINVAL;

	_cpu_mask_set_gen_inc(set);

	/* Find out CPUs left in CPU mask */
	cpumask_andnot(diff, &set->mask, &set->used);

	cpu = cpumask_first(diff);
	if (cpu >= nr_cpu_ids) /* empty */
		cpu = -EINVAL;
	else
		cpumask_set_cpu(cpu, &set->used);

	return cpu;
}

static void cpu_mask_set_put(struct cpu_mask_set *set, int cpu)
{
	if (!set)
		return;

	cpumask_clear_cpu(cpu, &set->used);
	_cpu_mask_set_gen_dec(set);
}

/* Initialize non-HT cpu cores mask */
void init_real_cpu_mask(void)
{
	int possible, curr_cpu, i, ht;

	cpumask_clear(&node_affinity.real_cpu_mask);

	/* Start with cpu online mask as the real cpu mask */
	cpumask_copy(&node_affinity.real_cpu_mask, cpu_online_mask);

	/*
	 * Remove HT cores from the real cpu mask.  Do this in two steps below.
	 */
	possible = cpumask_weight(&node_affinity.real_cpu_mask);
	ht = cpumask_weight(topology_sibling_cpumask(
				cpumask_first(&node_affinity.real_cpu_mask)));
	/*
	 * Step 1.  Skip over the first N HT siblings and use them as the
	 * "real" cores.  Assumes that HT cores are not enumerated in
	 * succession (except in the single core case).
	 */
	curr_cpu = cpumask_first(&node_affinity.real_cpu_mask);
	for (i = 0; i < possible / ht; i++)
		curr_cpu = cpumask_next(curr_cpu, &node_affinity.real_cpu_mask);
	/*
	 * Step 2.  Remove the remaining HT siblings.  Use cpumask_next() to
	 * skip any gaps.
	 */
	for (; i < possible; i++) {
		cpumask_clear_cpu(curr_cpu, &node_affinity.real_cpu_mask);
		curr_cpu = cpumask_next(curr_cpu, &node_affinity.real_cpu_mask);
	}
}

int node_affinity_init(void)
{
	int node;
	struct pci_dev *dev = NULL;
	const struct pci_device_id *ids = hfi1_pci_tbl;

	cpumask_clear(&node_affinity.proc.used);
	cpumask_copy(&node_affinity.proc.mask, cpu_online_mask);

	node_affinity.proc.gen = 0;
	node_affinity.num_core_siblings =
				cpumask_weight(topology_sibling_cpumask(
					cpumask_first(&node_affinity.proc.mask)
					));
	node_affinity.num_possible_nodes