[mirror_ubuntu-artful-kernel.git] / drivers / acpi / numa.c

/*
 *  acpi_numa.c - ACPI NUMA support
 *
 *  Copyright (C) 2002 Takayoshi Kochi <t-kochi@bq.jp.nec.com>
 *
 * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 *
 */

#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/bootmem.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;
}

/**
 * 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, n, dist, min_dist;

	node = acpi_map_pxm_to_node(pxm);

	if (node == NUMA_NO_NODE)
		node = 0;

	if (!node_online(node)) {
		min_dist = INT_MAX;
		for_each_online_node(n) {
			dist = node_distance(node, n);
			if (dist < min_dist) {
				min_dist = dist;
				node = n;
			}
		}
	}

	return 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%lx length 0x%lx) in proximity domain %d %s%s%s\n",
				 (unsigned long)p->base_address,
				 (unsigned 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(struct acpi_subtable_header *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);

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

	return 0;
}

static int __init
acpi_parse_processor_affinity(struct acpi_subtable_header *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);

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

	return 0;
}

static int __init
acpi_parse_gicc_affinity(struct acpi_subtable_header *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);

	/* 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(struct acpi_subtable_header * 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);

	/* 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,
					    NR_NODE_MEMBLKS);
	}

	/* 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
1da177e4 LT	1	/*
	2	* acpi_numa.c - ACPI NUMA support
	3	*
	4	* Copyright (C) 2002 Takayoshi Kochi <t-kochi@bq.jp.nec.com>
	5	*
	6	* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
	7	*
	8	* This program is free software; you can redistribute it and/or modify
	9	* it under the terms of the GNU General Public License as published by
	10	* the Free Software Foundation; either version 2 of the License, or
	11	* (at your option) any later version.
	12	*
	13	* This program is distributed in the hope that it will be useful,
	14	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	15	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	16	* GNU General Public License for more details.
	17	*
1da177e4 LT	18	* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
	19	*
	20	*/
ac906a6d HG	21
	22	#define pr_fmt(fmt) "ACPI: " fmt
	23
1da177e4	24	#include <linux/module.h>
1da177e4 LT	25	#include <linux/init.h>
	26	#include <linux/kernel.h>
	27	#include <linux/types.h>
	28	#include <linux/errno.h>
	29	#include <linux/acpi.h>
3770442e HG	30	#include <linux/bootmem.h>
3770442e HG	31	#include <linux/memblock.h>
b552a8c5	32	#include <linux/numa.h>
99759869 TK	33	#include <linux/nodemask.h>
99759869 TK	34	#include <linux/topology.h>
1da177e4	35
762834e8	36	static nodemask_t nodes_found_map = NODE_MASK_NONE;
762834e8 YG	37
762834e8 YG	38	/* maps to convert between proximity domain and logical node ID */
ffada891	39	static int pxm_to_node_map[MAX_PXM_DOMAINS]
b552a8c5	40	= { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
ffada891	41	static int node_to_pxm_map[MAX_NUMNODES]
b552a8c5	42	= { [0 ... MAX_NUMNODES - 1] = PXM_INVAL };
762834e8	43
8df0eb7c	44	unsigned char acpi_srat_revision __initdata;
e84025e2	45	int acpi_numa __initdata;
8df0eb7c	46
f363d16f	47	int pxm_to_node(int pxm)
762834e8 YG	48	{
762834e8 YG	49	if (pxm < 0)
b552a8c5	50	return NUMA_NO_NODE;
762834e8 YG	51	return pxm_to_node_map[pxm];
	52	}
	53
f363d16f	54	int node_to_pxm(int node)
762834e8 YG	55	{
	56	if (node < 0)
	57	return PXM_INVAL;
	58	return node_to_pxm_map[node];
	59	}
	60
d79ed248	61	static void __acpi_map_pxm_to_node(int pxm, int node)
3484d798	62	{
0f9b75ef DR	63	if (pxm_to_node_map[pxm] == NUMA_NO_NODE \|\| node < pxm_to_node_map[pxm])
	64	pxm_to_node_map[pxm] = node;
	65	if (node_to_pxm_map[node] == PXM_INVAL \|\| pxm < node_to_pxm_map[node])
	66	node_to_pxm_map[node] = pxm;
3484d798 DR	67	}
3484d798 DR	68
8ff6f48d	69	int acpi_map_pxm_to_node(int pxm)
762834e8	70	{
99759869 TK	71	int node;
99759869 TK	72
aec03f89	73	if (pxm < 0 \|\| pxm >= MAX_PXM_DOMAINS \|\| numa_off)
99759869 TK	74	return NUMA_NO_NODE;
	75
	76	node = pxm_to_node_map[pxm];
762834e8	77
1bb25df0	78	if (node == NUMA_NO_NODE) {
762834e8	79	if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
b552a8c5	80	return NUMA_NO_NODE;
762834e8	81	node = first_unset_node(nodes_found_map);
3484d798	82	__acpi_map_pxm_to_node(pxm, node);
762834e8 YG	83	node_set(node, nodes_found_map);
	84	}
	85
	86	return node;
	87	}
	88
99759869 TK	89	/**
	90	* acpi_map_pxm_to_online_node - Map proximity ID to online node
	91	* @pxm: ACPI proximity ID
	92	*
	93	* This is similar to acpi_map_pxm_to_node(), but always returns an online
	94	* node. When the mapped node from a given proximity ID is offline, it
	95	* looks up the node distance table and returns the nearest online node.
	96	*
	97	* ACPI device drivers, which are called after the NUMA initialization has
	98	* completed in the kernel, can call this interface to obtain their device
	99	* NUMA topology from ACPI tables. Such drivers do not have to deal with
	100	* offline nodes. A node may be offline when a device proximity ID is
	101	* unique, SRAT memory entry does not exist, or NUMA is disabled, ex.
	102	* "numa=off" on x86.
	103	*/
	104	int acpi_map_pxm_to_online_node(int pxm)
	105	{
	106	int node, n, dist, min_dist;
	107
	108	node = acpi_map_pxm_to_node(pxm);
	109
	110	if (node == NUMA_NO_NODE)
	111	node = 0;
	112
	113	if (!node_online(node)) {
	114	min_dist = INT_MAX;
	115	for_each_online_node(n) {
	116	dist = node_distance(node, n);
	117	if (dist < min_dist) {
	118	min_dist = dist;
	119	node = n;
	120	}
	121	}
	122	}
	123
	124	return node;
	125	}
	126	EXPORT_SYMBOL(acpi_map_pxm_to_online_node);
	127
ae2c6dcf DR	128	static void __init
ae2c6dcf DR	129	acpi_table_print_srat_entry(struct acpi_subtable_header *header)
1da177e4	130	{
1da177e4	131	switch (header->type) {
15a58ed1	132	case ACPI_SRAT_TYPE_CPU_AFFINITY:
4be44fcd	133	{
15a58ed1 AS	134	struct acpi_srat_cpu_affinity *p =
15a58ed1 AS	135	(struct acpi_srat_cpu_affinity *)header;
3dda4481 HG	136	pr_debug("SRAT Processor (id[0x%02x] eid[0x%02x]) in proximity domain %d %s\n",
	137	p->apic_id, p->local_sapic_eid,
	138	p->proximity_domain_lo,
	139	(p->flags & ACPI_SRAT_CPU_ENABLED) ?
	140	"enabled" : "disabled");
4be44fcd	141	}
1da177e4 LT	142	break;
1da177e4 LT	143
15a58ed1	144	case ACPI_SRAT_TYPE_MEMORY_AFFINITY:
4be44fcd	145	{
15a58ed1 AS	146	struct acpi_srat_mem_affinity *p =
15a58ed1 AS	147	(struct acpi_srat_mem_affinity *)header;
3dda4481 HG	148	pr_debug("SRAT Memory (0x%lx length 0x%lx) in proximity domain %d %s%s%s\n",
	149	(unsigned long)p->base_address,
	150	(unsigned long)p->length,
	151	p->proximity_domain,
	152	(p->flags & ACPI_SRAT_MEM_ENABLED) ?
	153	"enabled" : "disabled",
	154	(p->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE) ?
	155	" hot-pluggable" : "",
	156	(p->flags & ACPI_SRAT_MEM_NON_VOLATILE) ?
	157	" non-volatile" : "");
4be44fcd	158	}
1da177e4 LT	159	break;
1da177e4 LT	160
7237d3de	161	case ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY:
7237d3de SS	162	{
	163	struct acpi_srat_x2apic_cpu_affinity *p =
	164	(struct acpi_srat_x2apic_cpu_affinity *)header;
3dda4481 HG	165	pr_debug("SRAT Processor (x2apicid[0x%08x]) in proximity domain %d %s\n",
	166	p->apic_id,
	167	p->proximity_domain,
	168	(p->flags & ACPI_SRAT_CPU_ENABLED) ?
	169	"enabled" : "disabled");
7237d3de	170	}
7237d3de	171	break;
3dda4481	172
4bac6fa7 HG	173	case ACPI_SRAT_TYPE_GICC_AFFINITY:
	174	{
	175	struct acpi_srat_gicc_affinity *p =
	176	(struct acpi_srat_gicc_affinity *)header;
	177	pr_debug("SRAT Processor (acpi id[0x%04x]) in proximity domain %d %s\n",
	178	p->acpi_processor_uid,
	179	p->proximity_domain,
	180	(p->flags & ACPI_SRAT_GICC_ENABLED) ?
	181	"enabled" : "disabled");
	182	}
	183	break;
	184
1da177e4	185	default:
ac906a6d HG	186	pr_warn("Found unsupported SRAT entry (type = 0x%x)\n",
ac906a6d HG	187	header->type);
1da177e4 LT	188	break;
	189	}
	190	}
	191
39b8931b FY	192	/*
	193	* A lot of BIOS fill in 10 (= no distance) everywhere. This messes
	194	* up the NUMA heuristics which wants the local node to have a smaller
	195	* distance than the others.
	196	* Do some quick checks here and only use the SLIT if it passes.
	197	*/
40e31856	198	static int __init slit_valid(struct acpi_table_slit *slit)
39b8931b FY	199	{
	200	int i, j;
	201	int d = slit->locality_count;
	202	for (i = 0; i < d; i++) {
	203	for (j = 0; j < d; j++) {
	204	u8 val = slit->entry[d*i + j];
	205	if (i == j) {
	206	if (val != LOCAL_DISTANCE)
	207	return 0;
	208	} else if (val <= LOCAL_DISTANCE)
	209	return 0;
	210	}
	211	}
	212	return 1;
	213	}
	214
e84025e2 DD	215	void __init bad_srat(void)
	216	{
	217	pr_err("SRAT: SRAT not used.\n");
	218	acpi_numa = -1;
	219	}
	220
	221	int __init srat_disabled(void)
	222	{
	223	return acpi_numa < 0;
	224	}
	225
6525afdf HG	226	#if defined(CONFIG_X86) \|\| defined(CONFIG_ARM64)
	227	/*
	228	* Callback for SLIT parsing. pxm_to_node() returns NUMA_NO_NODE for
	229	* I/O localities since SRAT does not list them. I/O localities are
	230	* not supported at this point.
	231	*/
	232	void __init acpi_numa_slit_init(struct acpi_table_slit *slit)
	233	{
	234	int i, j;
	235
	236	for (i = 0; i < slit->locality_count; i++) {
	237	const int from_node = pxm_to_node(i);
	238
	239	if (from_node == NUMA_NO_NODE)
	240	continue;
	241
	242	for (j = 0; j < slit->locality_count; j++) {
	243	const int to_node = pxm_to_node(j);
	244
	245	if (to_node == NUMA_NO_NODE)
	246	continue;
	247
	248	numa_set_distance(from_node, to_node,
	249	slit->entry[slit->locality_count * i + j]);
	250	}
	251	}
	252	}
3770442e HG	253
	254	/*
	255	* Default callback for parsing of the Proximity Domain <-> Memory
	256	* Area mappings
	257	*/
	258	int __init
	259	acpi_numa_memory_affinity_init(struct acpi_srat_mem_affinity *ma)
	260	{
	261	u64 start, end;
	262	u32 hotpluggable;
	263	int node, pxm;
	264
	265	if (srat_disabled())
	266	goto out_err;
e0af261a DD	267	if (ma->header.length < sizeof(struct acpi_srat_mem_affinity)) {
	268	pr_err("SRAT: Unexpected header length: %d\n",
	269	ma->header.length);
3770442e	270	goto out_err_bad_srat;
e0af261a	271	}
3770442e HG	272	if ((ma->flags & ACPI_SRAT_MEM_ENABLED) == 0)
	273	goto out_err;
	274	hotpluggable = ma->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE;
	275	if (hotpluggable && !IS_ENABLED(CONFIG_MEMORY_HOTPLUG))
	276	goto out_err;
	277
	278	start = ma->base_address;
	279	end = start + ma->length;
	280	pxm = ma->proximity_domain;
	281	if (acpi_srat_revision <= 1)
	282	pxm &= 0xff;
	283
	284	node = acpi_map_pxm_to_node(pxm);
e0af261a DD	285	if (node == NUMA_NO_NODE \|\| node >= MAX_NUMNODES) {
e0af261a DD	286	pr_err("SRAT: Too many proximity domains.\n");
3770442e HG	287	goto out_err_bad_srat;
	288	}
	289
e0af261a DD	290	if (numa_add_memblk(node, start, end) < 0) {
	291	pr_err("SRAT: Failed to add memblk to node %u [mem %#010Lx-%#010Lx]\n",
	292	node, (unsigned long long) start,
	293	(unsigned long long) end - 1);
3770442e	294	goto out_err_bad_srat;
e0af261a	295	}
3770442e HG	296
	297	node_set(node, numa_nodes_parsed);
	298
	299	pr_info("SRAT: Node %u PXM %u [mem %#010Lx-%#010Lx]%s%s\n",
	300	node, pxm,
	301	(unsigned long long) start, (unsigned long long) end - 1,
	302	hotpluggable ? " hotplug" : "",
	303	ma->flags & ACPI_SRAT_MEM_NON_VOLATILE ? " non-volatile" : "");
	304
	305	/* Mark hotplug range in memblock. */
	306	if (hotpluggable && memblock_mark_hotplug(start, ma->length))
	307	pr_warn("SRAT: Failed to mark hotplug range [mem %#010Lx-%#010Lx] in memblock\n",
	308	(unsigned long long)start, (unsigned long long)end - 1);
	309
	310	max_possible_pfn = max(max_possible_pfn, PFN_UP(end - 1));
	311
	312	return 0;
	313	out_err_bad_srat:
	314	bad_srat();
	315	out_err:
	316	return -EINVAL;
	317	}
6525afdf HG	318	#endif /* defined(CONFIG_X86) \|\| defined (CONFIG_ARM64) */
6525afdf HG	319
15a58ed1	320	static int __init acpi_parse_slit(struct acpi_table_header *table)
1da177e4	321	{
2fad9308	322	struct acpi_table_slit slit = (struct acpi_table_slit )table;
1da177e4	323
39b8931b	324	if (!slit_valid(slit)) {
ac906a6d	325	pr_info("SLIT table looks invalid. Not used.\n");
39b8931b FY	326	return -EINVAL;
39b8931b FY	327	}
1da177e4 LT	328	acpi_numa_slit_init(slit);
	329
	330	return 0;
	331	}
	332
beffbe54	333	void __init __weak
7237d3de SS	334	acpi_numa_x2apic_affinity_init(struct acpi_srat_x2apic_cpu_affinity *pa)
7237d3de SS	335	{
ac906a6d	336	pr_warn("Found unsupported x2apic [0x%08x] SRAT entry\n", pa->apic_id);
7237d3de SS	337	}
7237d3de SS	338
7237d3de SS	339	static int __init
	340	acpi_parse_x2apic_affinity(struct acpi_subtable_header *header,
	341	const unsigned long end)
	342	{
	343	struct acpi_srat_x2apic_cpu_affinity *processor_affinity;
	344
	345	processor_affinity = (struct acpi_srat_x2apic_cpu_affinity *)header;
	346	if (!processor_affinity)
	347	return -EINVAL;
	348
	349	acpi_table_print_srat_entry(header);
	350
	351	/* let architecture-dependent part to do it */
	352	acpi_numa_x2apic_affinity_init(processor_affinity);
	353
	354	return 0;
	355	}
	356
1da177e4	357	static int __init
7237d3de	358	acpi_parse_processor_affinity(struct acpi_subtable_header *header,
4be44fcd	359	const unsigned long end)
1da177e4	360	{
15a58ed1	361	struct acpi_srat_cpu_affinity *processor_affinity;
1da177e4	362
15a58ed1	363	processor_affinity = (struct acpi_srat_cpu_affinity *)header;
1da177e4 LT	364	if (!processor_affinity)
	365	return -EINVAL;
	366
	367	acpi_table_print_srat_entry(header);
	368
	369	/* let architecture-dependent part to do it */
	370	acpi_numa_processor_affinity_init(processor_affinity);
	371
	372	return 0;
	373	}
	374
4bac6fa7 HG	375	static int __init
	376	acpi_parse_gicc_affinity(struct acpi_subtable_header *header,
	377	const unsigned long end)
	378	{
	379	struct acpi_srat_gicc_affinity *processor_affinity;
	380
	381	processor_affinity = (struct acpi_srat_gicc_affinity *)header;
	382	if (!processor_affinity)
	383	return -EINVAL;
	384
	385	acpi_table_print_srat_entry(header);
	386
	387	/* let architecture-dependent part to do it */
	388	acpi_numa_gicc_affinity_init(processor_affinity);
	389
	390	return 0;
	391	}
	392
095adbb6 TR	393	static int __initdata parsed_numa_memblks;
095adbb6 TR	394
1da177e4	395	static int __init
15a58ed1	396	acpi_parse_memory_affinity(struct acpi_subtable_header * header,
4be44fcd	397	const unsigned long end)
1da177e4	398	{
15a58ed1	399	struct acpi_srat_mem_affinity *memory_affinity;
1da177e4	400
15a58ed1	401	memory_affinity = (struct acpi_srat_mem_affinity *)header;
1da177e4 LT	402	if (!memory_affinity)
	403	return -EINVAL;
	404
	405	acpi_table_print_srat_entry(header);
	406
	407	/* let architecture-dependent part to do it */
095adbb6 TR	408	if (!acpi_numa_memory_affinity_init(memory_affinity))
095adbb6 TR	409	parsed_numa_memblks++;
1da177e4 LT	410	return 0;
	411	}
	412
15a58ed1	413	static int __init acpi_parse_srat(struct acpi_table_header *table)
1da177e4	414	{
2fad9308	415	struct acpi_table_srat srat = (struct acpi_table_srat )table;
1da177e4	416
8df0eb7c KG	417	acpi_srat_revision = srat->header.revision;
8df0eb7c KG	418
cfa806f0	419	/* Real work done in acpi_table_parse_srat below. */
1da177e4 LT	420
	421	return 0;
	422	}
	423
ae2c6dcf	424	static int __init
15a58ed1	425	acpi_table_parse_srat(enum acpi_srat_type id,
b43e1065	426	acpi_tbl_entry_handler handler, unsigned int max_entries)
1da177e4	427	{
6eb87fed	428	return acpi_table_parse_entries(ACPI_SIG_SRAT,
4be44fcd LB	429	sizeof(struct acpi_table_srat), id,
4be44fcd LB	430	handler, max_entries);
1da177e4 LT	431	}
1da177e4 LT	432
20e6926d	433	int __init acpi_numa_init(void)
e8d19552	434	{
20e6926d YL	435	int cnt = 0;
20e6926d YL	436
4bac6fa7 HG	437	if (acpi_disabled)
	438	return -EINVAL;
	439
d3bd0588 YL	440	/*
	441	* Should not limit number with cpu num that is from NR_CPUS or nr_cpus=
	442	* SRAT cpu entries could have different order with that in MADT.
	443	* So go over all cpu entries in SRAT to get apicid to node mapping.
	444	*/
8716273c	445
02cb489b	446	/* SRAT: System Resource Affinity Table */
7f8f97c3	447	if (!acpi_table_parse(ACPI_SIG_SRAT, acpi_parse_srat)) {
4bac6fa7	448	struct acpi_subtable_proc srat_proc[3];
702b07fc LA	449
	450	memset(srat_proc, 0, sizeof(srat_proc));
	451	srat_proc[0].id = ACPI_SRAT_TYPE_CPU_AFFINITY;
	452	srat_proc[0].handler = acpi_parse_processor_affinity;
	453	srat_proc[1].id = ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY;
	454	srat_proc[1].handler = acpi_parse_x2apic_affinity;
4bac6fa7 HG	455	srat_proc[2].id = ACPI_SRAT_TYPE_GICC_AFFINITY;
4bac6fa7 HG	456	srat_proc[2].handler = acpi_parse_gicc_affinity;
702b07fc LA	457
	458	acpi_table_parse_entries_array(ACPI_SIG_SRAT,
	459	sizeof(struct acpi_table_srat),
	460	srat_proc, ARRAY_SIZE(srat_proc), 0);
	461
20e6926d YL	462	cnt = acpi_table_parse_srat(ACPI_SRAT_TYPE_MEMORY_AFFINITY,
	463	acpi_parse_memory_affinity,
	464	NR_NODE_MEMBLKS);
1da177e4 LT	465	}
	466
	467	/* SLIT: System Locality Information Table */
7f8f97c3	468	acpi_table_parse(ACPI_SIG_SLIT, acpi_parse_slit);
1da177e4	469
20e6926d YL	470	if (cnt < 0)
20e6926d YL	471	return cnt;
095adbb6	472	else if (!parsed_numa_memblks)
f3946fb6	473	return -ENOENT;
940fed2e	474	return 0;
1da177e4 LT	475	}
1da177e4 LT	476
d79ed248	477	static int acpi_get_pxm(acpi_handle h)
1da177e4	478	{
27663c58	479	unsigned long long pxm;
1da177e4 LT	480	acpi_status status;
	481	acpi_handle handle;
	482	acpi_handle phandle = h;
	483
	484	do {
	485	handle = phandle;
	486	status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm);
	487	if (ACPI_SUCCESS(status))
50dd0969	488	return pxm;
1da177e4	489	status = acpi_get_parent(handle, &phandle);
4be44fcd	490	} while (ACPI_SUCCESS(status));
1da177e4 LT	491	return -1;
1da177e4 LT	492	}
1e3590e2	493
486c79b5	494	int acpi_get_node(acpi_handle handle)
1e3590e2	495	{
962fe9c9	496	int pxm;
1e3590e2	497
1e3590e2	498	pxm = acpi_get_pxm(handle);
1e3590e2	499
962fe9c9	500	return acpi_map_pxm_to_node(pxm);
1e3590e2 YG	501	}
1e3590e2 YG	502	EXPORT_SYMBOL(acpi_get_node);