[linux-block.git] / drivers / acpi / numa.c

// SPDX-License-Identifier: GPL-2.0-or-later
/*
 *  acpi_numa.c - ACPI NUMA support
 *
 *  Copyright (C) 2002 Takayoshi Kochi <t-kochi@bq.jp.nec.com>
 */

#define pr_fmt(fmt) "ACPI: " fmt

#include <linux/module.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/errno.h>
#include <linux/acpi.h>
#include <linux/memblock.h>
#include <linux/numa.h>
#include <linux/nodemask.h>
#include <linux/topology.h>

static nodemask_t nodes_found_map = NODE_MASK_NONE;

/* maps to convert between proximity domain and logical node ID */
static int pxm_to_node_map[MAX_PXM_DOMAINS]
			= { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
static int node_to_pxm_map[MAX_NUMNODES]
			= { [0 ... MAX_NUMNODES - 1] = PXM_INVAL };

unsigned char acpi_srat_revision __initdata;
int acpi_numa __initdata;

int pxm_to_node(int pxm)
{
	if (pxm < 0)
		return NUMA_NO_NODE;
	return pxm_to_node_map[pxm];
}

int node_to_pxm(int node)
{
	if (node < 0)
		return PXM_INVAL;
	return node_to_pxm_map[node];
}

static void __acpi_map_pxm_to_node(int pxm, int node)
{
	if (pxm_to_node_map[pxm] == NUMA_NO_NODE || node < pxm_to_node_map[pxm])
		pxm_to_node_map[pxm] = node;
	if (node_to_pxm_map[node] == PXM_INVAL || pxm < node_to_pxm_map[node])
		node_to_pxm_map[node] = pxm;
}

int acpi_map_pxm_to_node(int pxm)
{
	int node;

	if (pxm < 0 || pxm >= MAX_PXM_DOMAINS || numa_off)
		return NUMA_NO_NODE;

	node = pxm_to_node_map[pxm];

	if (node == NUMA_NO_NODE) {
		if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
			return NUMA_NO_NODE;
		node = first_unset_node(nodes_found_map);
		__acpi_map_pxm_to_node(pxm, node);
		node_set(node, nodes_found_map);
	}

	return node;
}
EXPORT_SYMBOL(acpi_map_pxm_to_node);

/**
 * acpi_map_pxm_to_online_node - Map proximity ID to online node
 * @pxm: ACPI proximity ID
 *
 * This is similar to acpi_map_pxm_to_node(), but always returns an online
 * node.  When the mapped node from a given proximity ID is offline, it
 * looks up the node distance table and returns the nearest online node.
 *
 * ACPI device drivers, which are called after the NUMA initialization has
 * completed in the kernel, can call this interface to obtain their device
 * NUMA topology from ACPI tables.  Such drivers do not have to deal with
 * offline nodes.  A node may be offline when a device proximity ID is
 * unique, SRAT memory entry does not exist, or NUMA is disabled, ex.
 * "numa=off" on x86.
 */
int acpi_map_pxm_to_online_node(int pxm)
{
	int node, min_node;

	node = acpi_map_pxm_to_node(pxm);

	if (node == NUMA_NO_NODE)
		node = 0;

	min_node = node;
	if (!node_online(node)) {
		int min_dist = INT_MAX, dist, n;

		for_each_online_node(n) {
			dist = node_distance(node, n);
			if (dist < min_dist) {
				min_dist = dist;
				min_node = n;
			}
		}
	}

	return min_node;
}
EXPORT_SYMBOL(acpi_map_pxm_to_online_node);

static void __init
acpi_table_print_srat_entry(struct acpi_subtable_header *header)
{
	switch (header->type) {
	case ACPI_SRAT_TYPE_CPU_AFFINITY:
		{
			struct acpi_srat_cpu_affinity *p =
			    (struct acpi_srat_cpu_affinity *)header;
			pr_debug("SRAT Processor (id[0x%02x] eid[0x%02x]) in proximity domain %d %s\n",
				 p->apic_id, p->local_sapic_eid,
				 p->proximity_domain_lo,
				 (p->flags & ACPI_SRAT_CPU_ENABLED) ?
				 "enabled" : "disabled");
		}
		break;

	case ACPI_SRAT_TYPE_MEMORY_AFFINITY:
		{
			struct acpi_srat_mem_affinity *p =
			    (struct acpi_srat_mem_affinity *)header;
			pr_debug("SRAT Memory (0x%llx length 0x%llx) in proximity domain %d %s%s%s\n",
				 (unsigned long long)p->base_address,
				 (unsigned long long)p->length,
				 p->proximity_domain,
				 (p->flags & ACPI_SRAT_MEM_ENABLED) ?
				 "enabled" : "disabled",
				 (p->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE) ?
				 " hot-pluggable" : "",
				 (p->flags & ACPI_SRAT_MEM_NON_VOLATILE) ?
				 " non-volatile" : "");
		}
		break;

	case ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY:
		{
			struct acpi_srat_x2apic_cpu_affinity *p =
			    (struct acpi_srat_x2apic_cpu_affinity *)header;
			pr_debug("SRAT Processor (x2apicid[0x%08x]) in proximity domain %d %s\n",
				 p->apic_id,
				 p->proximity_domain,
				 (p->flags & ACPI_SRAT_CPU_ENABLED) ?
				 "enabled" : "disabled");
		}
		break;

	case ACPI_SRAT_TYPE_GICC_AFFINITY:
		{
			struct acpi_srat_gicc_affinity *p =
			    (struct acpi_srat_gicc_affinity *)header;
			pr_debug("SRAT Processor (acpi id[0x%04x]) in proximity domain %d %s\n",
				 p->acpi_processor_uid,
				 p->proximity_domain,
				 (p->flags & ACPI_SRAT_GICC_ENABLED) ?
				 "enabled" : "disabled");
		}
		break;

	default:
		pr_warn("Found unsupported SRAT entry (type = 0x%x)\n",
			header->type);
		break;
	}
}

/*
 * A lot of BIOS fill in 10 (= no distance) everywhere. This messes
 * up the NUMA heuristics which wants the local node to have a smaller
 * distance than the others.
 * Do some quick checks here and only use the SLIT if it passes.
 */
static int __init slit_valid(struct acpi_table_slit *slit)
{
	int i, j;
	int d = slit->locality_count;
	for (i = 0; i < d; i++) {
		for (j = 0; j < d; j++)  {
			u8 val = slit->entry[d*i + j];
			if (i == j) {
				if (val != LOCAL_DISTANCE)
					return 0;
			} else if (val <= LOCAL_DISTANCE)
				return 0;
		}
	}
	return 1;
}

void __init bad_srat(void)
{
	pr_err("SRAT: SRAT not used.\n");
	acpi_numa = -1;
}

int __init srat_disabled(void)
{
	return acpi_numa < 0;
}

#if defined(CONFIG_X86) || defined(CONFIG_ARM64)
/*
 * Callback for SLIT parsing.  pxm_to_node() returns NUMA_NO_NODE for
 * I/O localities since SRAT does not list them.  I/O localities are
 * not supported at this point.
 */
void __init acpi_numa_slit_init(struct acpi_table_slit *slit)
{
	int i, j;

	for (i = 0; i < slit->locality_count; i++) {
		const int from_node = pxm_to_node(i);

		if (from_node == NUMA_NO_NODE)
			continue;

		for (j = 0; j < slit->locality_count; j++) {
			const int to_node = pxm_to_node(j);

			if (to_node == NUMA_NO_NODE)
				continue;

			numa_set_distance(from_node, to_node,
				slit->entry[slit->locality_count * i + j]);
		}
	}
}

/*
 * Default callback for parsing of the Proximity Domain <-> Memory
 * Area mappings
 */
int __init
acpi_numa_memory_affinity_init(struct acpi_srat_mem_affinity *ma)
{
	u64 start, end;
	u32 hotpluggable;
	int node, pxm;

	if (srat_disabled())
		goto out_err;
	if (ma->header.length < sizeof(struct acpi_srat_mem_affinity)) {
		pr_err("SRAT: Unexpected header length: %d\n",
		       ma->header.length);
		goto out_err_bad_srat;
	}
	if ((ma->flags & ACPI_SRAT_MEM_ENABLED) == 0)
		goto out_err;
	hotpluggable = ma->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE;
	if (hotpluggable && !IS_ENABLED(CONFIG_MEMORY_HOTPLUG))
		goto out_err;

	start = ma->base_address;
	end = start + ma->length;
	pxm = ma->proximity_domain;
	if (acpi_srat_revision <= 1)
		pxm &= 0xff;

	node = acpi_map_pxm_to_node(pxm);
	if (node == NUMA_NO_NODE || node >= MAX_NUMNODES) {
		pr_err("SRAT: Too many proximity domains.\n");
		goto out_err_bad_srat;
	}

	if (numa_add_memblk(node, start, end) < 0) {
		pr_err("SRAT: Failed to add memblk to node %u [mem %#010Lx-%#010Lx]\n",
		       node, (unsigned long long) start,
		       (unsigned long long) end - 1);
		goto out_err_bad_srat;
	}

	node_set(node, numa_nodes_parsed);

	pr_info("SRAT: Node %u PXM %u [mem %#010Lx-%#010Lx]%s%s\n",
		node, pxm,
		(unsigned long long) start, (unsigned long long) end - 1,
		hotpluggable ? " hotplug" : "",
		ma->flags & ACPI_SRAT_MEM_NON_VOLATILE ? " non-volatile" : "");

	/* Mark hotplug range in memblock. */
	if (hotpluggable && memblock_mark_hotplug(start, ma->length))
		pr_warn("SRAT: Failed to mark hotplug range [mem %#010Lx-%#010Lx] in memblock\n",
			(unsigned long long)start, (unsigned long long)end - 1);

	max_possible_pfn = max(max_possible_pfn, PFN_UP(end - 1));

	return 0;
out_err_bad_srat:
	bad_srat();
out_err:
	return -EINVAL;
}
#endif /* defined(CONFIG_X86) || defined (CONFIG_ARM64) */

static int __init acpi_parse_slit(struct acpi_table_header *table)
{
	struct acpi_table_slit *slit = (struct acpi_table_slit *)table;

	if (!slit_valid(slit)) {
		pr_info("SLIT table looks invalid. Not used.\n");
		return -EINVAL;
	}
	acpi_numa_slit_init(slit);

	return 0;
}

void __init __weak
acpi_numa_x2apic_affinity_init(struct acpi_srat_x2apic_cpu_affinity *pa)
{
	pr_warn("Found unsupported x2apic [0x%08x] SRAT entry\n", pa->apic_id);
}

static int __init
acpi_parse_x2apic_affinity(union acpi_subtable_headers *header,
			   const unsigned long end)
{
	struct acpi_srat_x2apic_cpu_affinity *processor_affinity;

	processor_affinity = (struct acpi_srat_x2apic_cpu_affinity *)header;
	if (!processor_affinity)
		return -EINVAL;

	acpi_table_print_srat_entry(&header->common);

	/* let architecture-dependent part to do it */
	acpi_numa_x2apic_affinity_init(processor_affinity);

	return 0;
}

static int __init
acpi_parse_processor_affinity(union acpi_subtable_headers *header,
			      const unsigned long end)
{
	struct acpi_srat_cpu_affinity *processor_affinity;

	processor_affinity = (struct acpi_srat_cpu_affinity *)header;
	if (!processor_affinity)
		return -EINVAL;

	acpi_table_print_srat_entry(&header->common);

	/* let architecture-dependent part to do it */
	acpi_numa_processor_affinity_init(processor_affinity);

	return 0;
}

static int __init
acpi_parse_gicc_affinity(union acpi_subtable_headers *header,
			 const unsigned long end)
{
	struct acpi_srat_gicc_affinity *processor_affinity;

	processor_affinity = (struct acpi_srat_gicc_affinity *)header;
	if (!processor_affinity)
		return -EINVAL;

	acpi_table_print_srat_entry(&header->common);

	/* let architecture-dependent part to do it */
	acpi_numa_gicc_affinity_init(processor_affinity);

	return 0;
}

static int __initdata parsed_numa_memblks;

static int __init
acpi_parse_memory_affinity(union acpi_subtable_headers * header,
			   const unsigned long end)
{
	struct acpi_srat_mem_affinity *memory_affinity;

	memory_affinity = (struct acpi_srat_mem_affinity *)header;
	if (!memory_affinity)
		return -EINVAL;

	acpi_table_print_srat_entry(&header->common);

	/* let architecture-dependent part to do it */
	if (!acpi_numa_memory_affinity_init(memory_affinity))
		parsed_numa_memblks++;
	return 0;
}

static int __init acpi_parse_srat(struct acpi_table_header *table)
{
	struct acpi_table_srat *srat = (struct acpi_table_srat *)table;

	acpi_srat_revision = srat->header.revision;

	/* Real work done in acpi_table_parse_srat below. */

	return 0;
}

static int __init
acpi_table_parse_srat(enum acpi_srat_type id,
		      acpi_tbl_entry_handler handler, unsigned int max_entries)
{
	return acpi_table_parse_entries(ACPI_SIG_SRAT,
					    sizeof(struct acpi_table_srat), id,
					    handler, max_entries);
}

int __init acpi_numa_init(void)
{
	int cnt = 0;

	if (acpi_disabled)
		return -EINVAL;

	/*
	 * Should not limit number with cpu num that is from NR_CPUS or nr_cpus=
	 * SRAT cpu entries could have different order with that in MADT.
	 * So go over all cpu entries in SRAT to get apicid to node mapping.
	 */

	/* SRAT: System Resource Affinity Table */
	if (!acpi_table_parse(ACPI_SIG_SRAT, acpi_parse_srat)) {
		struct acpi_subtable_proc srat_proc[3];

		memset(srat_proc, 0, sizeof(srat_proc));
		srat_proc[0].id = ACPI_SRAT_TYPE_CPU_AFFINITY;
		srat_proc[0].handler = acpi_parse_processor_affinity;
		srat_proc[1].id = ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY;
		srat_proc[1].handler = acpi_parse_x2apic_affinity;
		srat_proc[2].id = ACPI_SRAT_TYPE_GICC_AFFINITY;
		srat_proc[2].handler = acpi_parse_gicc_affinity;

		acpi_table_parse_entries_array(ACPI_SIG_SRAT,
					sizeof(struct acpi_table_srat),
					srat_proc, ARRAY_SIZE(srat_proc), 0);

		cnt = acpi_table_parse_srat(ACPI_SRAT_TYPE_MEMORY_AFFINITY,
					    acpi_parse_memory_affinity, 0);
	}

	/* SLIT: System Locality Information Table */
	acpi_table_parse(ACPI_SIG_SLIT, acpi_parse_slit);

	if (cnt < 0)
		return cnt;
	else if (!parsed_numa_memblks)
		return -ENOENT;
	return 0;
}

static int acpi_get_pxm(acpi_handle h)
{
	unsigned long long pxm;
	acpi_status status;
	acpi_handle handle;
	acpi_handle phandle = h;

	do {
		handle = phandle;
		status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm);
		if (ACPI_SUCCESS(status))
			return pxm;
		status = acpi_get_parent(handle, &phandle);
	} while (ACPI_SUCCESS(status));
	return -1;
}

int acpi_get_node(acpi_handle handle)
{
	int pxm;

	pxm = acpi_get_pxm(handle);

	return acpi_map_pxm_to_node(pxm);
}
EXPORT_SYMBOL(acpi_get_node);
Commit	Line	Data
c942fddf	1	// SPDX-License-Identifier: GPL-2.0-or-later
1da177e4 LT	2	/*
	3	* acpi_numa.c - ACPI NUMA support
	4	*
	5	* Copyright (C) 2002 Takayoshi Kochi <t-kochi@bq.jp.nec.com>
1da177e4	6	*/
ac906a6d HG	7
	8	#define pr_fmt(fmt) "ACPI: " fmt
	9
1da177e4	10	#include <linux/module.h>
1da177e4 LT	11	#include <linux/init.h>
	12	#include <linux/kernel.h>
	13	#include <linux/types.h>
	14	#include <linux/errno.h>
	15	#include <linux/acpi.h>
3770442e	16	#include <linux/memblock.h>
b552a8c5	17	#include <linux/numa.h>
99759869 TK	18	#include <linux/nodemask.h>
99759869 TK	19	#include <linux/topology.h>
1da177e4	20
762834e8	21	static nodemask_t nodes_found_map = NODE_MASK_NONE;
762834e8 YG	22
762834e8 YG	23	/* maps to convert between proximity domain and logical node ID */
ffada891	24	static int pxm_to_node_map[MAX_PXM_DOMAINS]
b552a8c5	25	= { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
ffada891	26	static int node_to_pxm_map[MAX_NUMNODES]
b552a8c5	27	= { [0 ... MAX_NUMNODES - 1] = PXM_INVAL };
762834e8	28
8df0eb7c	29	unsigned char acpi_srat_revision __initdata;
e84025e2	30	int acpi_numa __initdata;
8df0eb7c	31
f363d16f	32	int pxm_to_node(int pxm)
762834e8 YG	33	{
762834e8 YG	34	if (pxm < 0)
b552a8c5	35	return NUMA_NO_NODE;
762834e8 YG	36	return pxm_to_node_map[pxm];
	37	}
	38
f363d16f	39	int node_to_pxm(int node)
762834e8 YG	40	{
	41	if (node < 0)
	42	return PXM_INVAL;
	43	return node_to_pxm_map[node];
	44	}
	45
d79ed248	46	static void __acpi_map_pxm_to_node(int pxm, int node)
3484d798	47	{
0f9b75ef DR	48	if (pxm_to_node_map[pxm] == NUMA_NO_NODE \|\| node < pxm_to_node_map[pxm])
	49	pxm_to_node_map[pxm] = node;
	50	if (node_to_pxm_map[node] == PXM_INVAL \|\| pxm < node_to_pxm_map[node])
	51	node_to_pxm_map[node] = pxm;
3484d798 DR	52	}
3484d798 DR	53
8ff6f48d	54	int acpi_map_pxm_to_node(int pxm)
762834e8	55	{
99759869 TK	56	int node;
99759869 TK	57
aec03f89	58	if (pxm < 0 \|\| pxm >= MAX_PXM_DOMAINS \|\| numa_off)
99759869 TK	59	return NUMA_NO_NODE;
	60
	61	node = pxm_to_node_map[pxm];
762834e8	62
1bb25df0	63	if (node == NUMA_NO_NODE) {
762834e8	64	if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
b552a8c5	65	return NUMA_NO_NODE;
762834e8	66	node = first_unset_node(nodes_found_map);
3484d798	67	__acpi_map_pxm_to_node(pxm, node);
762834e8 YG	68	node_set(node, nodes_found_map);
	69	}
	70
	71	return node;
	72	}
8fc5c735	73	EXPORT_SYMBOL(acpi_map_pxm_to_node);
762834e8	74
99759869 TK	75	/**
	76	* acpi_map_pxm_to_online_node - Map proximity ID to online node
	77	* @pxm: ACPI proximity ID
	78	*
	79	* This is similar to acpi_map_pxm_to_node(), but always returns an online
	80	* node. When the mapped node from a given proximity ID is offline, it
	81	* looks up the node distance table and returns the nearest online node.
	82	*
	83	* ACPI device drivers, which are called after the NUMA initialization has
	84	* completed in the kernel, can call this interface to obtain their device
	85	* NUMA topology from ACPI tables. Such drivers do not have to deal with
	86	* offline nodes. A node may be offline when a device proximity ID is
	87	* unique, SRAT memory entry does not exist, or NUMA is disabled, ex.
	88	* "numa=off" on x86.
	89	*/
	90	int acpi_map_pxm_to_online_node(int pxm)
	91	{
dc9e0a93	92	int node, min_node;
99759869 TK	93
	94	node = acpi_map_pxm_to_node(pxm);
	95
	96	if (node == NUMA_NO_NODE)
	97	node = 0;
	98
dc9e0a93	99	min_node = node;
99759869	100	if (!node_online(node)) {
dc9e0a93 DW	101	int min_dist = INT_MAX, dist, n;
dc9e0a93 DW	102
99759869 TK	103	for_each_online_node(n) {
	104	dist = node_distance(node, n);
	105	if (dist < min_dist) {
	106	min_dist = dist;
dc9e0a93	107	min_node = n;
99759869 TK	108	}
	109	}
	110	}
	111
dc9e0a93	112	return min_node;
99759869 TK	113	}
	114	EXPORT_SYMBOL(acpi_map_pxm_to_online_node);
	115
ae2c6dcf DR	116	static void __init
ae2c6dcf DR	117	acpi_table_print_srat_entry(struct acpi_subtable_header *header)
1da177e4	118	{
1da177e4	119	switch (header->type) {
15a58ed1	120	case ACPI_SRAT_TYPE_CPU_AFFINITY:
4be44fcd	121	{
15a58ed1 AS	122	struct acpi_srat_cpu_affinity *p =
15a58ed1 AS	123	(struct acpi_srat_cpu_affinity *)header;
3dda4481 HG	124	pr_debug("SRAT Processor (id[0x%02x] eid[0x%02x]) in proximity domain %d %s\n",
	125	p->apic_id, p->local_sapic_eid,
	126	p->proximity_domain_lo,
	127	(p->flags & ACPI_SRAT_CPU_ENABLED) ?
	128	"enabled" : "disabled");
4be44fcd	129	}
1da177e4 LT	130	break;
1da177e4 LT	131
15a58ed1	132	case ACPI_SRAT_TYPE_MEMORY_AFFINITY:
4be44fcd	133	{
15a58ed1 AS	134	struct acpi_srat_mem_affinity *p =
15a58ed1 AS	135	(struct acpi_srat_mem_affinity *)header;
b9ced18a CF	136	pr_debug("SRAT Memory (0x%llx length 0x%llx) in proximity domain %d %s%s%s\n",
	137	(unsigned long long)p->base_address,
	138	(unsigned long long)p->length,
3dda4481 HG	139	p->proximity_domain,
	140	(p->flags & ACPI_SRAT_MEM_ENABLED) ?
	141	"enabled" : "disabled",
	142	(p->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE) ?
	143	" hot-pluggable" : "",
	144	(p->flags & ACPI_SRAT_MEM_NON_VOLATILE) ?
	145	" non-volatile" : "");
4be44fcd	146	}
1da177e4 LT	147	break;
1da177e4 LT	148
7237d3de	149	case ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY:
7237d3de SS	150	{
	151	struct acpi_srat_x2apic_cpu_affinity *p =
	152	(struct acpi_srat_x2apic_cpu_affinity *)header;
3dda4481 HG	153	pr_debug("SRAT Processor (x2apicid[0x%08x]) in proximity domain %d %s\n",
	154	p->apic_id,
	155	p->proximity_domain,
	156	(p->flags & ACPI_SRAT_CPU_ENABLED) ?
	157	"enabled" : "disabled");
7237d3de	158	}
7237d3de	159	break;
3dda4481	160
4bac6fa7 HG	161	case ACPI_SRAT_TYPE_GICC_AFFINITY:
	162	{
	163	struct acpi_srat_gicc_affinity *p =
	164	(struct acpi_srat_gicc_affinity *)header;
	165	pr_debug("SRAT Processor (acpi id[0x%04x]) in proximity domain %d %s\n",
	166	p->acpi_processor_uid,
	167	p->proximity_domain,
	168	(p->flags & ACPI_SRAT_GICC_ENABLED) ?
	169	"enabled" : "disabled");
	170	}
	171	break;
	172
1da177e4	173	default:
ac906a6d HG	174	pr_warn("Found unsupported SRAT entry (type = 0x%x)\n",
ac906a6d HG	175	header->type);
1da177e4 LT	176	break;
	177	}
	178	}
	179
39b8931b FY	180	/*
	181	* A lot of BIOS fill in 10 (= no distance) everywhere. This messes
	182	* up the NUMA heuristics which wants the local node to have a smaller
	183	* distance than the others.
	184	* Do some quick checks here and only use the SLIT if it passes.
	185	*/
40e31856	186	static int __init slit_valid(struct acpi_table_slit *slit)
39b8931b FY	187	{
	188	int i, j;
	189	int d = slit->locality_count;
	190	for (i = 0; i < d; i++) {
	191	for (j = 0; j < d; j++) {
	192	u8 val = slit->entry[d*i + j];
	193	if (i == j) {
	194	if (val != LOCAL_DISTANCE)
	195	return 0;
	196	} else if (val <= LOCAL_DISTANCE)
	197	return 0;
	198	}
	199	}
	200	return 1;
	201	}
	202
e84025e2 DD	203	void __init bad_srat(void)
	204	{
	205	pr_err("SRAT: SRAT not used.\n");
	206	acpi_numa = -1;
	207	}
	208
	209	int __init srat_disabled(void)
	210	{
	211	return acpi_numa < 0;
	212	}
	213
6525afdf HG	214	#if defined(CONFIG_X86) \|\| defined(CONFIG_ARM64)
	215	/*
	216	* Callback for SLIT parsing. pxm_to_node() returns NUMA_NO_NODE for
	217	* I/O localities since SRAT does not list them. I/O localities are
	218	* not supported at this point.
	219	*/
	220	void __init acpi_numa_slit_init(struct acpi_table_slit *slit)
	221	{
	222	int i, j;
	223
	224	for (i = 0; i < slit->locality_count; i++) {
	225	const int from_node = pxm_to_node(i);
	226
	227	if (from_node == NUMA_NO_NODE)
	228	continue;
	229
	230	for (j = 0; j < slit->locality_count; j++) {
	231	const int to_node = pxm_to_node(j);
	232
	233	if (to_node == NUMA_NO_NODE)
	234	continue;
	235
	236	numa_set_distance(from_node, to_node,
	237	slit->entry[slit->locality_count * i + j]);
	238	}
	239	}
	240	}
3770442e HG	241
	242	/*
	243	* Default callback for parsing of the Proximity Domain <-> Memory
	244	* Area mappings
	245	*/
	246	int __init
	247	acpi_numa_memory_affinity_init(struct acpi_srat_mem_affinity *ma)
	248	{
	249	u64 start, end;
	250	u32 hotpluggable;
	251	int node, pxm;
	252
	253	if (srat_disabled())
	254	goto out_err;
e0af261a DD	255	if (ma->header.length < sizeof(struct acpi_srat_mem_affinity)) {
	256	pr_err("SRAT: Unexpected header length: %d\n",
	257	ma->header.length);
3770442e	258	goto out_err_bad_srat;
e0af261a	259	}
3770442e HG	260	if ((ma->flags & ACPI_SRAT_MEM_ENABLED) == 0)
	261	goto out_err;
	262	hotpluggable = ma->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE;
	263	if (hotpluggable && !IS_ENABLED(CONFIG_MEMORY_HOTPLUG))
	264	goto out_err;
	265
	266	start = ma->base_address;
	267	end = start + ma->length;
	268	pxm = ma->proximity_domain;
	269	if (acpi_srat_revision <= 1)
	270	pxm &= 0xff;
	271
	272	node = acpi_map_pxm_to_node(pxm);
e0af261a DD	273	if (node == NUMA_NO_NODE \|\| node >= MAX_NUMNODES) {
e0af261a DD	274	pr_err("SRAT: Too many proximity domains.\n");
3770442e HG	275	goto out_err_bad_srat;
	276	}
	277
e0af261a DD	278	if (numa_add_memblk(node, start, end) < 0) {
	279	pr_err("SRAT: Failed to add memblk to node %u [mem %#010Lx-%#010Lx]\n",
	280	node, (unsigned long long) start,
	281	(unsigned long long) end - 1);
3770442e	282	goto out_err_bad_srat;
e0af261a	283	}
3770442e HG	284
	285	node_set(node, numa_nodes_parsed);
	286
	287	pr_info("SRAT: Node %u PXM %u [mem %#010Lx-%#010Lx]%s%s\n",
	288	node, pxm,
	289	(unsigned long long) start, (unsigned long long) end - 1,
	290	hotpluggable ? " hotplug" : "",
	291	ma->flags & ACPI_SRAT_MEM_NON_VOLATILE ? " non-volatile" : "");
	292
	293	/* Mark hotplug range in memblock. */
	294	if (hotpluggable && memblock_mark_hotplug(start, ma->length))
	295	pr_warn("SRAT: Failed to mark hotplug range [mem %#010Lx-%#010Lx] in memblock\n",
	296	(unsigned long long)start, (unsigned long long)end - 1);
	297
	298	max_possible_pfn = max(max_possible_pfn, PFN_UP(end - 1));
	299
	300	return 0;
	301	out_err_bad_srat:
	302	bad_srat();
	303	out_err:
	304	return -EINVAL;
	305	}
6525afdf HG	306	#endif /* defined(CONFIG_X86) \|\| defined (CONFIG_ARM64) */
6525afdf HG	307
15a58ed1	308	static int __init acpi_parse_slit(struct acpi_table_header *table)
1da177e4	309	{
2fad9308	310	struct acpi_table_slit slit = (struct acpi_table_slit )table;
1da177e4	311
39b8931b	312	if (!slit_valid(slit)) {
ac906a6d	313	pr_info("SLIT table looks invalid. Not used.\n");
39b8931b FY	314	return -EINVAL;
39b8931b FY	315	}
1da177e4 LT	316	acpi_numa_slit_init(slit);
	317
	318	return 0;
	319	}
	320
beffbe54	321	void __init __weak
7237d3de SS	322	acpi_numa_x2apic_affinity_init(struct acpi_srat_x2apic_cpu_affinity *pa)
7237d3de SS	323	{
ac906a6d	324	pr_warn("Found unsupported x2apic [0x%08x] SRAT entry\n", pa->apic_id);
7237d3de SS	325	}
7237d3de SS	326
7237d3de	327	static int __init
60574d1e	328	acpi_parse_x2apic_affinity(union acpi_subtable_headers *header,
7237d3de SS	329	const unsigned long end)
	330	{
	331	struct acpi_srat_x2apic_cpu_affinity *processor_affinity;
	332
	333	processor_affinity = (struct acpi_srat_x2apic_cpu_affinity *)header;
	334	if (!processor_affinity)
	335	return -EINVAL;
	336
60574d1e	337	acpi_table_print_srat_entry(&header->common);
7237d3de SS	338
	339	/* let architecture-dependent part to do it */
	340	acpi_numa_x2apic_affinity_init(processor_affinity);
	341
	342	return 0;
	343	}
	344
1da177e4	345	static int __init
60574d1e	346	acpi_parse_processor_affinity(union acpi_subtable_headers *header,
4be44fcd	347	const unsigned long end)
1da177e4	348	{
15a58ed1	349	struct acpi_srat_cpu_affinity *processor_affinity;
1da177e4	350
15a58ed1	351	processor_affinity = (struct acpi_srat_cpu_affinity *)header;
1da177e4 LT	352	if (!processor_affinity)
	353	return -EINVAL;
	354
60574d1e	355	acpi_table_print_srat_entry(&header->common);
1da177e4 LT	356
	357	/* let architecture-dependent part to do it */
	358	acpi_numa_processor_affinity_init(processor_affinity);
	359
	360	return 0;
	361	}
	362
4bac6fa7	363	static int __init
60574d1e	364	acpi_parse_gicc_affinity(union acpi_subtable_headers *header,
4bac6fa7 HG	365	const unsigned long end)
	366	{
	367	struct acpi_srat_gicc_affinity *processor_affinity;
	368
	369	processor_affinity = (struct acpi_srat_gicc_affinity *)header;
	370	if (!processor_affinity)
	371	return -EINVAL;
	372
60574d1e	373	acpi_table_print_srat_entry(&header->common);
4bac6fa7 HG	374
	375	/* let architecture-dependent part to do it */
	376	acpi_numa_gicc_affinity_init(processor_affinity);
	377
	378	return 0;
	379	}
	380
095adbb6 TR	381	static int __initdata parsed_numa_memblks;
095adbb6 TR	382
1da177e4	383	static int __init
60574d1e	384	acpi_parse_memory_affinity(union acpi_subtable_headers * header,
4be44fcd	385	const unsigned long end)
1da177e4	386	{
15a58ed1	387	struct acpi_srat_mem_affinity *memory_affinity;
1da177e4	388
15a58ed1	389	memory_affinity = (struct acpi_srat_mem_affinity *)header;
1da177e4 LT	390	if (!memory_affinity)
	391	return -EINVAL;
	392
60574d1e	393	acpi_table_print_srat_entry(&header->common);
1da177e4 LT	394
1da177e4 LT	395	/* let architecture-dependent part to do it */
095adbb6 TR	396	if (!acpi_numa_memory_affinity_init(memory_affinity))
095adbb6 TR	397	parsed_numa_memblks++;
1da177e4 LT	398	return 0;
	399	}
	400
15a58ed1	401	static int __init acpi_parse_srat(struct acpi_table_header *table)
1da177e4	402	{
2fad9308	403	struct acpi_table_srat srat = (struct acpi_table_srat )table;
1da177e4	404
8df0eb7c KG	405	acpi_srat_revision = srat->header.revision;
8df0eb7c KG	406
cfa806f0	407	/* Real work done in acpi_table_parse_srat below. */
1da177e4 LT	408
	409	return 0;
	410	}
	411
ae2c6dcf	412	static int __init
15a58ed1	413	acpi_table_parse_srat(enum acpi_srat_type id,
b43e1065	414	acpi_tbl_entry_handler handler, unsigned int max_entries)
1da177e4	415	{
6eb87fed	416	return acpi_table_parse_entries(ACPI_SIG_SRAT,
4be44fcd LB	417	sizeof(struct acpi_table_srat), id,
4be44fcd LB	418	handler, max_entries);
1da177e4 LT	419	}
1da177e4 LT	420
20e6926d	421	int __init acpi_numa_init(void)
e8d19552	422	{
20e6926d YL	423	int cnt = 0;
20e6926d YL	424
4bac6fa7 HG	425	if (acpi_disabled)
	426	return -EINVAL;
	427
d3bd0588 YL	428	/*
	429	* Should not limit number with cpu num that is from NR_CPUS or nr_cpus=
	430	* SRAT cpu entries could have different order with that in MADT.
	431	* So go over all cpu entries in SRAT to get apicid to node mapping.
	432	*/
8716273c	433
02cb489b	434	/* SRAT: System Resource Affinity Table */
7f8f97c3	435	if (!acpi_table_parse(ACPI_SIG_SRAT, acpi_parse_srat)) {
4bac6fa7	436	struct acpi_subtable_proc srat_proc[3];
702b07fc LA	437
	438	memset(srat_proc, 0, sizeof(srat_proc));
	439	srat_proc[0].id = ACPI_SRAT_TYPE_CPU_AFFINITY;
	440	srat_proc[0].handler = acpi_parse_processor_affinity;
	441	srat_proc[1].id = ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY;
	442	srat_proc[1].handler = acpi_parse_x2apic_affinity;
4bac6fa7 HG	443	srat_proc[2].id = ACPI_SRAT_TYPE_GICC_AFFINITY;
4bac6fa7 HG	444	srat_proc[2].handler = acpi_parse_gicc_affinity;
702b07fc LA	445
	446	acpi_table_parse_entries_array(ACPI_SIG_SRAT,
	447	sizeof(struct acpi_table_srat),
	448	srat_proc, ARRAY_SIZE(srat_proc), 0);
	449
20e6926d	450	cnt = acpi_table_parse_srat(ACPI_SRAT_TYPE_MEMORY_AFFINITY,
fd3e4543	451	acpi_parse_memory_affinity, 0);
1da177e4 LT	452	}
	453
	454	/* SLIT: System Locality Information Table */
7f8f97c3	455	acpi_table_parse(ACPI_SIG_SLIT, acpi_parse_slit);
1da177e4	456
20e6926d YL	457	if (cnt < 0)
20e6926d YL	458	return cnt;
095adbb6	459	else if (!parsed_numa_memblks)
f3946fb6	460	return -ENOENT;
940fed2e	461	return 0;
1da177e4 LT	462	}
1da177e4 LT	463
d79ed248	464	static int acpi_get_pxm(acpi_handle h)
1da177e4	465	{
27663c58	466	unsigned long long pxm;
1da177e4 LT	467	acpi_status status;
	468	acpi_handle handle;
	469	acpi_handle phandle = h;
	470
	471	do {
	472	handle = phandle;
	473	status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm);
	474	if (ACPI_SUCCESS(status))
50dd0969	475	return pxm;
1da177e4	476	status = acpi_get_parent(handle, &phandle);
4be44fcd	477	} while (ACPI_SUCCESS(status));
1da177e4 LT	478	return -1;
1da177e4 LT	479	}
1e3590e2	480
486c79b5	481	int acpi_get_node(acpi_handle handle)
1e3590e2	482	{
962fe9c9	483	int pxm;
1e3590e2	484
1e3590e2	485	pxm = acpi_get_pxm(handle);
1e3590e2	486
962fe9c9	487	return acpi_map_pxm_to_node(pxm);
1e3590e2 YG	488	}
1e3590e2 YG	489	EXPORT_SYMBOL(acpi_get_node);