2005-04-17 02:20:36 +04:00
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program; if not, write to the Free Software
|
|
|
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
|
|
|
*
|
|
|
|
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/acpi.h>
|
2009-11-18 01:27:32 +03:00
|
|
|
#include <linux/numa.h>
|
2005-04-17 02:20:36 +04:00
|
|
|
#include <acpi/acpi_bus.h>
|
|
|
|
|
2009-07-29 00:45:54 +04:00
|
|
|
#define PREFIX "ACPI: "
|
|
|
|
|
2005-04-17 02:20:36 +04:00
|
|
|
#define ACPI_NUMA 0x80000000
|
|
|
|
#define _COMPONENT ACPI_NUMA
|
2007-02-13 06:42:12 +03:00
|
|
|
ACPI_MODULE_NAME("numa");
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2006-06-23 13:03:19 +04:00
|
|
|
static nodemask_t nodes_found_map = NODE_MASK_NONE;
|
|
|
|
|
|
|
|
/* maps to convert between proximity domain and logical node ID */
|
2007-12-13 11:32:26 +03:00
|
|
|
static int pxm_to_node_map[MAX_PXM_DOMAINS]
|
2009-11-18 01:27:32 +03:00
|
|
|
= { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
|
2007-12-13 11:32:26 +03:00
|
|
|
static int node_to_pxm_map[MAX_NUMNODES]
|
2009-11-18 01:27:32 +03:00
|
|
|
= { [0 ... MAX_NUMNODES - 1] = PXM_INVAL };
|
2006-06-23 13:03:19 +04:00
|
|
|
|
2012-01-17 13:18:02 +04:00
|
|
|
unsigned char acpi_srat_revision __initdata;
|
|
|
|
|
2007-05-17 09:11:06 +04:00
|
|
|
int pxm_to_node(int pxm)
|
2006-06-23 13:03:19 +04:00
|
|
|
{
|
|
|
|
if (pxm < 0)
|
2009-11-18 01:27:32 +03:00
|
|
|
return NUMA_NO_NODE;
|
2006-06-23 13:03:19 +04:00
|
|
|
return pxm_to_node_map[pxm];
|
|
|
|
}
|
|
|
|
|
2007-05-17 09:11:06 +04:00
|
|
|
int node_to_pxm(int node)
|
2006-06-23 13:03:19 +04:00
|
|
|
{
|
|
|
|
if (node < 0)
|
|
|
|
return PXM_INVAL;
|
|
|
|
return node_to_pxm_map[node];
|
|
|
|
}
|
|
|
|
|
2007-07-21 19:10:32 +04:00
|
|
|
void __acpi_map_pxm_to_node(int pxm, int node)
|
|
|
|
{
|
2010-03-26 02:33:04 +03:00
|
|
|
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;
|
2007-07-21 19:10:32 +04:00
|
|
|
}
|
|
|
|
|
2007-05-25 00:57:40 +04:00
|
|
|
int acpi_map_pxm_to_node(int pxm)
|
2006-06-23 13:03:19 +04:00
|
|
|
{
|
|
|
|
int node = pxm_to_node_map[pxm];
|
|
|
|
|
2009-11-18 01:27:32 +03:00
|
|
|
if (node < 0) {
|
2006-06-23 13:03:19 +04:00
|
|
|
if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
|
2009-11-18 01:27:32 +03:00
|
|
|
return NUMA_NO_NODE;
|
2006-06-23 13:03:19 +04:00
|
|
|
node = first_unset_node(nodes_found_map);
|
2007-07-21 19:10:32 +04:00
|
|
|
__acpi_map_pxm_to_node(pxm, node);
|
2006-06-23 13:03:19 +04:00
|
|
|
node_set(node, nodes_found_map);
|
|
|
|
}
|
|
|
|
|
|
|
|
return node;
|
|
|
|
}
|
|
|
|
|
2007-07-21 19:09:56 +04:00
|
|
|
static void __init
|
|
|
|
acpi_table_print_srat_entry(struct acpi_subtable_header *header)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
|
|
|
|
2005-08-05 08:44:28 +04:00
|
|
|
ACPI_FUNCTION_NAME("acpi_table_print_srat_entry");
|
2005-04-17 02:20:36 +04:00
|
|
|
|
|
|
|
if (!header)
|
|
|
|
return;
|
|
|
|
|
|
|
|
switch (header->type) {
|
|
|
|
|
2007-02-02 19:48:22 +03:00
|
|
|
case ACPI_SRAT_TYPE_CPU_AFFINITY:
|
2005-04-17 02:20:36 +04:00
|
|
|
#ifdef ACPI_DEBUG_OUTPUT
|
2005-08-05 08:44:28 +04:00
|
|
|
{
|
2007-02-02 19:48:22 +03:00
|
|
|
struct acpi_srat_cpu_affinity *p =
|
|
|
|
(struct acpi_srat_cpu_affinity *)header;
|
2005-08-05 08:44:28 +04:00
|
|
|
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
|
|
|
|
"SRAT Processor (id[0x%02x] eid[0x%02x]) in proximity domain %d %s\n",
|
2007-02-02 19:48:22 +03:00
|
|
|
p->apic_id, p->local_sapic_eid,
|
|
|
|
p->proximity_domain_lo,
|
|
|
|
(p->flags & ACPI_SRAT_CPU_ENABLED)?
|
|
|
|
"enabled" : "disabled"));
|
2005-08-05 08:44:28 +04:00
|
|
|
}
|
|
|
|
#endif /* ACPI_DEBUG_OUTPUT */
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
|
2007-02-02 19:48:22 +03:00
|
|
|
case ACPI_SRAT_TYPE_MEMORY_AFFINITY:
|
2005-04-17 02:20:36 +04:00
|
|
|
#ifdef ACPI_DEBUG_OUTPUT
|
2005-08-05 08:44:28 +04:00
|
|
|
{
|
2007-02-02 19:48:22 +03:00
|
|
|
struct acpi_srat_mem_affinity *p =
|
|
|
|
(struct acpi_srat_mem_affinity *)header;
|
2005-08-05 08:44:28 +04:00
|
|
|
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
|
2013-01-10 02:06:08 +04:00
|
|
|
"SRAT Memory (0x%lx length 0x%lx) in proximity domain %d %s%s%s\n",
|
2007-02-02 19:48:22 +03:00
|
|
|
(unsigned long)p->base_address,
|
|
|
|
(unsigned long)p->length,
|
2008-06-10 11:54:40 +04:00
|
|
|
p->proximity_domain,
|
2007-02-02 19:48:22 +03:00
|
|
|
(p->flags & ACPI_SRAT_MEM_ENABLED)?
|
|
|
|
"enabled" : "disabled",
|
|
|
|
(p->flags & ACPI_SRAT_MEM_HOT_PLUGGABLE)?
|
2013-01-10 02:06:08 +04:00
|
|
|
" hot-pluggable" : "",
|
|
|
|
(p->flags & ACPI_SRAT_MEM_NON_VOLATILE)?
|
|
|
|
" non-volatile" : ""));
|
2005-08-05 08:44:28 +04:00
|
|
|
}
|
|
|
|
#endif /* ACPI_DEBUG_OUTPUT */
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
|
2009-03-31 01:55:30 +04:00
|
|
|
case ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY:
|
|
|
|
#ifdef ACPI_DEBUG_OUTPUT
|
|
|
|
{
|
|
|
|
struct acpi_srat_x2apic_cpu_affinity *p =
|
|
|
|
(struct acpi_srat_x2apic_cpu_affinity *)header;
|
|
|
|
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
|
|
|
|
"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"));
|
|
|
|
}
|
|
|
|
#endif /* ACPI_DEBUG_OUTPUT */
|
|
|
|
break;
|
2005-04-17 02:20:36 +04:00
|
|
|
default:
|
2005-08-05 08:44:28 +04:00
|
|
|
printk(KERN_WARNING PREFIX
|
|
|
|
"Found unsupported SRAT entry (type = 0x%x)\n",
|
|
|
|
header->type);
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-06-10 03:48:18 +04:00
|
|
|
/*
|
|
|
|
* 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 __init int 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;
|
|
|
|
}
|
|
|
|
|
2007-02-02 19:48:22 +03:00
|
|
|
static int __init acpi_parse_slit(struct acpi_table_header *table)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2005-08-05 08:44:28 +04:00
|
|
|
struct acpi_table_slit *slit;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2007-02-02 19:48:22 +03:00
|
|
|
if (!table)
|
2005-04-17 02:20:36 +04:00
|
|
|
return -EINVAL;
|
|
|
|
|
2007-02-02 19:48:22 +03:00
|
|
|
slit = (struct acpi_table_slit *)table;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2008-06-10 03:48:18 +04:00
|
|
|
if (!slit_valid(slit)) {
|
|
|
|
printk(KERN_INFO "ACPI: SLIT table looks invalid. Not used.\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
acpi_numa_slit_init(slit);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-03-31 01:55:30 +04:00
|
|
|
void __init __attribute__ ((weak))
|
|
|
|
acpi_numa_x2apic_affinity_init(struct acpi_srat_x2apic_cpu_affinity *pa)
|
|
|
|
{
|
|
|
|
printk(KERN_WARNING PREFIX
|
|
|
|
"Found unsupported x2apic [0x%08x] SRAT entry\n", pa->apic_id);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
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;
|
|
|
|
}
|
|
|
|
|
2005-04-17 02:20:36 +04:00
|
|
|
static int __init
|
2009-03-31 01:55:30 +04:00
|
|
|
acpi_parse_processor_affinity(struct acpi_subtable_header *header,
|
2005-08-05 08:44:28 +04:00
|
|
|
const unsigned long end)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2007-02-02 19:48:22 +03:00
|
|
|
struct acpi_srat_cpu_affinity *processor_affinity;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2007-02-02 19:48:22 +03:00
|
|
|
processor_affinity = (struct acpi_srat_cpu_affinity *)header;
|
2005-04-17 02:20:36 +04:00
|
|
|
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;
|
|
|
|
}
|
|
|
|
|
2012-07-31 19:41:09 +04:00
|
|
|
static int __initdata parsed_numa_memblks;
|
|
|
|
|
2005-04-17 02:20:36 +04:00
|
|
|
static int __init
|
2007-02-02 19:48:22 +03:00
|
|
|
acpi_parse_memory_affinity(struct acpi_subtable_header * header,
|
2005-08-05 08:44:28 +04:00
|
|
|
const unsigned long end)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2007-02-02 19:48:22 +03:00
|
|
|
struct acpi_srat_mem_affinity *memory_affinity;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2007-02-02 19:48:22 +03:00
|
|
|
memory_affinity = (struct acpi_srat_mem_affinity *)header;
|
2005-04-17 02:20:36 +04:00
|
|
|
if (!memory_affinity)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
acpi_table_print_srat_entry(header);
|
|
|
|
|
|
|
|
/* let architecture-dependent part to do it */
|
2012-07-31 19:41:09 +04:00
|
|
|
if (!acpi_numa_memory_affinity_init(memory_affinity))
|
|
|
|
parsed_numa_memblks++;
|
2005-04-17 02:20:36 +04:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-02-02 19:48:22 +03:00
|
|
|
static int __init acpi_parse_srat(struct acpi_table_header *table)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2012-01-17 13:18:02 +04:00
|
|
|
struct acpi_table_srat *srat;
|
2007-02-02 19:48:22 +03:00
|
|
|
if (!table)
|
2005-04-17 02:20:36 +04:00
|
|
|
return -EINVAL;
|
|
|
|
|
2012-01-17 13:18:02 +04:00
|
|
|
srat = (struct acpi_table_srat *)table;
|
|
|
|
acpi_srat_revision = srat->header.revision;
|
|
|
|
|
2010-07-21 02:18:36 +04:00
|
|
|
/* Real work done in acpi_table_parse_srat below. */
|
2005-04-17 02:20:36 +04:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-07-21 19:09:56 +04:00
|
|
|
static int __init
|
2007-02-02 19:48:22 +03:00
|
|
|
acpi_table_parse_srat(enum acpi_srat_type id,
|
2013-01-12 19:29:38 +04:00
|
|
|
acpi_tbl_entry_handler handler, unsigned int max_entries)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2007-02-11 06:17:07 +03:00
|
|
|
return acpi_table_parse_entries(ACPI_SIG_SRAT,
|
2005-08-05 08:44:28 +04:00
|
|
|
sizeof(struct acpi_table_srat), id,
|
|
|
|
handler, max_entries);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
acpi, memory-hotplug: parse SRAT before memblock is ready
On linux, the pages used by kernel could not be migrated. As a result,
if a memory range is used by kernel, it cannot be hot-removed. So if we
want to hot-remove memory, we should prevent kernel from using it.
The way now used to prevent this is specify a memory range by
movablemem_map boot option and set it as ZONE_MOVABLE.
But when the system is booting, memblock will allocate memory, and
reserve the memory for kernel. And before we parse SRAT, and know the
node memory ranges, memblock is working. And it may allocate memory in
ranges to be set as ZONE_MOVABLE. This memory can be used by kernel,
and never be freed.
So, let's parse SRAT before memblock is called first. And it is early
enough.
The first call of memblock_find_in_range_node() is in:
setup_arch()
|-->setup_real_mode()
so, this patch add a function early_parse_srat() to parse SRAT, and call
it before setup_real_mode() is called.
NOTE:
1) early_parse_srat() is called before numa_init(), and has initialized
numa_meminfo. So DO NOT clear numa_nodes_parsed in numa_init() and DO
NOT zero numa_meminfo in numa_init(), otherwise we will lose memory
numa info.
2) I don't know why using count of memory affinities parsed from SRAT
as a return value in original acpi_numa_init(). So I add a static
variable srat_mem_cnt to remember this count and use it as the return
value of the new acpi_numa_init()
[mhocko@suse.cz: parse SRAT before memblock is ready fix]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Wu Jianguo <wujianguo@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Len Brown <lenb@kernel.org>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 04:33:44 +04:00
|
|
|
static int srat_mem_cnt;
|
2010-12-17 06:09:58 +03:00
|
|
|
|
acpi, memory-hotplug: parse SRAT before memblock is ready
On linux, the pages used by kernel could not be migrated. As a result,
if a memory range is used by kernel, it cannot be hot-removed. So if we
want to hot-remove memory, we should prevent kernel from using it.
The way now used to prevent this is specify a memory range by
movablemem_map boot option and set it as ZONE_MOVABLE.
But when the system is booting, memblock will allocate memory, and
reserve the memory for kernel. And before we parse SRAT, and know the
node memory ranges, memblock is working. And it may allocate memory in
ranges to be set as ZONE_MOVABLE. This memory can be used by kernel,
and never be freed.
So, let's parse SRAT before memblock is called first. And it is early
enough.
The first call of memblock_find_in_range_node() is in:
setup_arch()
|-->setup_real_mode()
so, this patch add a function early_parse_srat() to parse SRAT, and call
it before setup_real_mode() is called.
NOTE:
1) early_parse_srat() is called before numa_init(), and has initialized
numa_meminfo. So DO NOT clear numa_nodes_parsed in numa_init() and DO
NOT zero numa_meminfo in numa_init(), otherwise we will lose memory
numa info.
2) I don't know why using count of memory affinities parsed from SRAT
as a return value in original acpi_numa_init(). So I add a static
variable srat_mem_cnt to remember this count and use it as the return
value of the new acpi_numa_init()
[mhocko@suse.cz: parse SRAT before memblock is ready fix]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Wu Jianguo <wujianguo@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Len Brown <lenb@kernel.org>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 04:33:44 +04:00
|
|
|
void __init early_parse_srat(void)
|
|
|
|
{
|
2010-12-17 06:09:58 +03:00
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*/
|
2009-09-26 02:20:04 +04:00
|
|
|
|
2005-04-17 02:20:36 +04:00
|
|
|
/* SRAT: Static Resource Affinity Table */
|
2007-02-11 05:28:03 +03:00
|
|
|
if (!acpi_table_parse(ACPI_SIG_SRAT, acpi_parse_srat)) {
|
2009-03-31 01:55:30 +04:00
|
|
|
acpi_table_parse_srat(ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY,
|
acpi, memory-hotplug: parse SRAT before memblock is ready
On linux, the pages used by kernel could not be migrated. As a result,
if a memory range is used by kernel, it cannot be hot-removed. So if we
want to hot-remove memory, we should prevent kernel from using it.
The way now used to prevent this is specify a memory range by
movablemem_map boot option and set it as ZONE_MOVABLE.
But when the system is booting, memblock will allocate memory, and
reserve the memory for kernel. And before we parse SRAT, and know the
node memory ranges, memblock is working. And it may allocate memory in
ranges to be set as ZONE_MOVABLE. This memory can be used by kernel,
and never be freed.
So, let's parse SRAT before memblock is called first. And it is early
enough.
The first call of memblock_find_in_range_node() is in:
setup_arch()
|-->setup_real_mode()
so, this patch add a function early_parse_srat() to parse SRAT, and call
it before setup_real_mode() is called.
NOTE:
1) early_parse_srat() is called before numa_init(), and has initialized
numa_meminfo. So DO NOT clear numa_nodes_parsed in numa_init() and DO
NOT zero numa_meminfo in numa_init(), otherwise we will lose memory
numa info.
2) I don't know why using count of memory affinities parsed from SRAT
as a return value in original acpi_numa_init(). So I add a static
variable srat_mem_cnt to remember this count and use it as the return
value of the new acpi_numa_init()
[mhocko@suse.cz: parse SRAT before memblock is ready fix]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Wu Jianguo <wujianguo@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Len Brown <lenb@kernel.org>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 04:33:44 +04:00
|
|
|
acpi_parse_x2apic_affinity, 0);
|
2007-07-21 19:09:56 +04:00
|
|
|
acpi_table_parse_srat(ACPI_SRAT_TYPE_CPU_AFFINITY,
|
acpi, memory-hotplug: parse SRAT before memblock is ready
On linux, the pages used by kernel could not be migrated. As a result,
if a memory range is used by kernel, it cannot be hot-removed. So if we
want to hot-remove memory, we should prevent kernel from using it.
The way now used to prevent this is specify a memory range by
movablemem_map boot option and set it as ZONE_MOVABLE.
But when the system is booting, memblock will allocate memory, and
reserve the memory for kernel. And before we parse SRAT, and know the
node memory ranges, memblock is working. And it may allocate memory in
ranges to be set as ZONE_MOVABLE. This memory can be used by kernel,
and never be freed.
So, let's parse SRAT before memblock is called first. And it is early
enough.
The first call of memblock_find_in_range_node() is in:
setup_arch()
|-->setup_real_mode()
so, this patch add a function early_parse_srat() to parse SRAT, and call
it before setup_real_mode() is called.
NOTE:
1) early_parse_srat() is called before numa_init(), and has initialized
numa_meminfo. So DO NOT clear numa_nodes_parsed in numa_init() and DO
NOT zero numa_meminfo in numa_init(), otherwise we will lose memory
numa info.
2) I don't know why using count of memory affinities parsed from SRAT
as a return value in original acpi_numa_init(). So I add a static
variable srat_mem_cnt to remember this count and use it as the return
value of the new acpi_numa_init()
[mhocko@suse.cz: parse SRAT before memblock is ready fix]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Wu Jianguo <wujianguo@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Len Brown <lenb@kernel.org>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 04:33:44 +04:00
|
|
|
acpi_parse_processor_affinity, 0);
|
|
|
|
srat_mem_cnt = acpi_table_parse_srat(ACPI_SRAT_TYPE_MEMORY_AFFINITY,
|
|
|
|
acpi_parse_memory_affinity,
|
|
|
|
NR_NODE_MEMBLKS);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
acpi, memory-hotplug: parse SRAT before memblock is ready
On linux, the pages used by kernel could not be migrated. As a result,
if a memory range is used by kernel, it cannot be hot-removed. So if we
want to hot-remove memory, we should prevent kernel from using it.
The way now used to prevent this is specify a memory range by
movablemem_map boot option and set it as ZONE_MOVABLE.
But when the system is booting, memblock will allocate memory, and
reserve the memory for kernel. And before we parse SRAT, and know the
node memory ranges, memblock is working. And it may allocate memory in
ranges to be set as ZONE_MOVABLE. This memory can be used by kernel,
and never be freed.
So, let's parse SRAT before memblock is called first. And it is early
enough.
The first call of memblock_find_in_range_node() is in:
setup_arch()
|-->setup_real_mode()
so, this patch add a function early_parse_srat() to parse SRAT, and call
it before setup_real_mode() is called.
NOTE:
1) early_parse_srat() is called before numa_init(), and has initialized
numa_meminfo. So DO NOT clear numa_nodes_parsed in numa_init() and DO
NOT zero numa_meminfo in numa_init(), otherwise we will lose memory
numa info.
2) I don't know why using count of memory affinities parsed from SRAT
as a return value in original acpi_numa_init(). So I add a static
variable srat_mem_cnt to remember this count and use it as the return
value of the new acpi_numa_init()
[mhocko@suse.cz: parse SRAT before memblock is ready fix]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Wu Jianguo <wujianguo@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Len Brown <lenb@kernel.org>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 04:33:44 +04:00
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
|
acpi, memory-hotplug: parse SRAT before memblock is ready
On linux, the pages used by kernel could not be migrated. As a result,
if a memory range is used by kernel, it cannot be hot-removed. So if we
want to hot-remove memory, we should prevent kernel from using it.
The way now used to prevent this is specify a memory range by
movablemem_map boot option and set it as ZONE_MOVABLE.
But when the system is booting, memblock will allocate memory, and
reserve the memory for kernel. And before we parse SRAT, and know the
node memory ranges, memblock is working. And it may allocate memory in
ranges to be set as ZONE_MOVABLE. This memory can be used by kernel,
and never be freed.
So, let's parse SRAT before memblock is called first. And it is early
enough.
The first call of memblock_find_in_range_node() is in:
setup_arch()
|-->setup_real_mode()
so, this patch add a function early_parse_srat() to parse SRAT, and call
it before setup_real_mode() is called.
NOTE:
1) early_parse_srat() is called before numa_init(), and has initialized
numa_meminfo. So DO NOT clear numa_nodes_parsed in numa_init() and DO
NOT zero numa_meminfo in numa_init(), otherwise we will lose memory
numa info.
2) I don't know why using count of memory affinities parsed from SRAT
as a return value in original acpi_numa_init(). So I add a static
variable srat_mem_cnt to remember this count and use it as the return
value of the new acpi_numa_init()
[mhocko@suse.cz: parse SRAT before memblock is ready fix]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Wu Jianguo <wujianguo@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Len Brown <lenb@kernel.org>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 04:33:44 +04:00
|
|
|
int __init acpi_numa_init(void)
|
|
|
|
{
|
2005-04-17 02:20:36 +04:00
|
|
|
/* SLIT: System Locality Information Table */
|
2007-02-11 05:28:03 +03:00
|
|
|
acpi_table_parse(ACPI_SIG_SLIT, acpi_parse_slit);
|
2005-04-17 02:20:36 +04:00
|
|
|
|
|
|
|
acpi_numa_arch_fixup();
|
2011-02-16 14:13:06 +03:00
|
|
|
|
acpi, memory-hotplug: parse SRAT before memblock is ready
On linux, the pages used by kernel could not be migrated. As a result,
if a memory range is used by kernel, it cannot be hot-removed. So if we
want to hot-remove memory, we should prevent kernel from using it.
The way now used to prevent this is specify a memory range by
movablemem_map boot option and set it as ZONE_MOVABLE.
But when the system is booting, memblock will allocate memory, and
reserve the memory for kernel. And before we parse SRAT, and know the
node memory ranges, memblock is working. And it may allocate memory in
ranges to be set as ZONE_MOVABLE. This memory can be used by kernel,
and never be freed.
So, let's parse SRAT before memblock is called first. And it is early
enough.
The first call of memblock_find_in_range_node() is in:
setup_arch()
|-->setup_real_mode()
so, this patch add a function early_parse_srat() to parse SRAT, and call
it before setup_real_mode() is called.
NOTE:
1) early_parse_srat() is called before numa_init(), and has initialized
numa_meminfo. So DO NOT clear numa_nodes_parsed in numa_init() and DO
NOT zero numa_meminfo in numa_init(), otherwise we will lose memory
numa info.
2) I don't know why using count of memory affinities parsed from SRAT
as a return value in original acpi_numa_init(). So I add a static
variable srat_mem_cnt to remember this count and use it as the return
value of the new acpi_numa_init()
[mhocko@suse.cz: parse SRAT before memblock is ready fix]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Wu Jianguo <wujianguo@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Len Brown <lenb@kernel.org>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 04:33:44 +04:00
|
|
|
if (srat_mem_cnt < 0)
|
|
|
|
return srat_mem_cnt;
|
2012-07-31 19:41:09 +04:00
|
|
|
else if (!parsed_numa_memblks)
|
2012-07-31 19:41:08 +04:00
|
|
|
return -ENOENT;
|
2011-02-16 14:13:06 +03:00
|
|
|
return 0;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
2007-05-09 07:27:01 +04:00
|
|
|
int acpi_get_pxm(acpi_handle h)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2008-10-10 10:22:59 +04:00
|
|
|
unsigned long long pxm;
|
2005-04-17 02:20:36 +04:00
|
|
|
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))
|
2006-10-01 02:28:50 +04:00
|
|
|
return pxm;
|
2005-04-17 02:20:36 +04:00
|
|
|
status = acpi_get_parent(handle, &phandle);
|
2005-08-05 08:44:28 +04:00
|
|
|
} while (ACPI_SUCCESS(status));
|
2005-04-17 02:20:36 +04:00
|
|
|
return -1;
|
|
|
|
}
|
2006-06-27 13:53:31 +04:00
|
|
|
|
2007-05-09 07:27:01 +04:00
|
|
|
int acpi_get_node(acpi_handle *handle)
|
2006-06-27 13:53:31 +04:00
|
|
|
{
|
|
|
|
int pxm, node = -1;
|
|
|
|
|
|
|
|
pxm = acpi_get_pxm(handle);
|
2009-03-04 22:55:29 +03:00
|
|
|
if (pxm >= 0 && pxm < MAX_PXM_DOMAINS)
|
2006-06-27 13:53:31 +04:00
|
|
|
node = acpi_map_pxm_to_node(pxm);
|
|
|
|
|
2006-07-02 00:48:23 +04:00
|
|
|
return node;
|
2006-06-27 13:53:31 +04:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(acpi_get_node);
|