2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/pci.h>
|
2008-07-03 04:50:26 +08:00
|
|
|
#include <linux/topology.h>
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
#include <linux/cpu.h>
|
2010-02-10 17:20:07 +08:00
|
|
|
#include <linux/range.h>
|
|
|
|
|
2011-01-11 00:20:23 +08:00
|
|
|
#include <asm/amd_nb.h>
|
2008-12-27 21:02:28 +08:00
|
|
|
#include <asm/pci_x86.h>
|
2008-06-13 02:19:23 +08:00
|
|
|
|
x86: get mp_bus_to_node early
Currently, on an amd k8 system with multi ht chains, the numa_node of
pci devices under /sys/devices/pci0000:80/* is always 0, even if that
chain is on node 1 or 2 or 3.
Workaround: pcibus_to_node(bus) is used when we want to get the node that
pci_device is on.
In struct device, we already have numa_node member, and we could use
dev_to_node()/set_dev_node() to get and set numa_node in the device.
set_dev_node is called in pci_device_add() with pcibus_to_node(bus),
and pcibus_to_node uses bus->sysdata for nodeid.
The problem is when pci_add_device is called, bus->sysdata is not assigned
correct nodeid yet. The result is that numa_node will always be 0.
pcibios_scan_root and pci_scan_root could take sysdata. So we need to get
mp_bus_to_node mapping before these two are called, and thus
get_mp_bus_to_node could get correct node for sysdata in root bus.
In scanning of the root bus, all child busses will take parent bus sysdata.
So all pci_device->dev.numa_node will be assigned correctly and automatically.
Later we could use dev_to_node(&pci_dev->dev) to get numa_node, and we
could also could make other bus specific device get the correct numa_node
too.
This is an updated version of pci_sysdata and Jeff's pci_domain patch.
[ mingo@elte.hu: build fix ]
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-19 19:20:09 +08:00
|
|
|
#include <asm/pci-direct.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2009-10-05 12:54:24 +08:00
|
|
|
#include "bus_numa.h"
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* This discovers the pcibus <-> node mapping on AMD K8.
|
2008-02-19 19:21:20 +08:00
|
|
|
* also get peer root bus resource for io,mmio
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
|
2008-02-19 19:21:20 +08:00
|
|
|
struct pci_hostbridge_probe {
|
|
|
|
u32 bus;
|
|
|
|
u32 slot;
|
|
|
|
u32 vendor;
|
|
|
|
u32 device;
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct pci_hostbridge_probe pci_probes[] __initdata = {
|
|
|
|
{ 0, 0x18, PCI_VENDOR_ID_AMD, 0x1100 },
|
|
|
|
{ 0, 0x18, PCI_VENDOR_ID_AMD, 0x1200 },
|
|
|
|
{ 0xff, 0, PCI_VENDOR_ID_AMD, 0x1200 },
|
|
|
|
{ 0, 0x18, PCI_VENDOR_ID_AMD, 0x1300 },
|
|
|
|
};
|
|
|
|
|
2010-02-10 17:20:07 +08:00
|
|
|
#define RANGE_NUM 16
|
|
|
|
|
2012-04-03 09:31:54 +08:00
|
|
|
static struct pci_root_info __init *find_pci_root_info(int node, int link)
|
|
|
|
{
|
|
|
|
struct pci_root_info *info;
|
|
|
|
|
|
|
|
/* find the position */
|
|
|
|
list_for_each_entry(info, &pci_root_infos, list)
|
|
|
|
if (info->node == node && info->link == link)
|
|
|
|
return info;
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2012-05-02 07:25:18 +08:00
|
|
|
static void __init set_mp_bus_range_to_node(int min_bus, int max_bus, int node)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_NUMA
|
|
|
|
int j;
|
|
|
|
|
|
|
|
for (j = min_bus; j <= max_bus; j++)
|
|
|
|
set_mp_bus_to_node(j, node);
|
|
|
|
#endif
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
/**
|
x86: get mp_bus_to_node early
Currently, on an amd k8 system with multi ht chains, the numa_node of
pci devices under /sys/devices/pci0000:80/* is always 0, even if that
chain is on node 1 or 2 or 3.
Workaround: pcibus_to_node(bus) is used when we want to get the node that
pci_device is on.
In struct device, we already have numa_node member, and we could use
dev_to_node()/set_dev_node() to get and set numa_node in the device.
set_dev_node is called in pci_device_add() with pcibus_to_node(bus),
and pcibus_to_node uses bus->sysdata for nodeid.
The problem is when pci_add_device is called, bus->sysdata is not assigned
correct nodeid yet. The result is that numa_node will always be 0.
pcibios_scan_root and pci_scan_root could take sysdata. So we need to get
mp_bus_to_node mapping before these two are called, and thus
get_mp_bus_to_node could get correct node for sysdata in root bus.
In scanning of the root bus, all child busses will take parent bus sysdata.
So all pci_device->dev.numa_node will be assigned correctly and automatically.
Later we could use dev_to_node(&pci_dev->dev) to get numa_node, and we
could also could make other bus specific device get the correct numa_node
too.
This is an updated version of pci_sysdata and Jeff's pci_domain patch.
[ mingo@elte.hu: build fix ]
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-19 19:20:09 +08:00
|
|
|
* early_fill_mp_bus_to_node()
|
|
|
|
* called before pcibios_scan_root and pci_scan_bus
|
2005-04-17 06:20:36 +08:00
|
|
|
* fills the mp_bus_to_cpumask array based according to the LDT Bus Number
|
|
|
|
* Registers found in the K8 northbridge
|
|
|
|
*/
|
2008-02-19 19:21:20 +08:00
|
|
|
static int __init early_fill_mp_bus_info(void)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2008-02-19 19:21:20 +08:00
|
|
|
int i;
|
|
|
|
unsigned bus;
|
x86: get mp_bus_to_node early
Currently, on an amd k8 system with multi ht chains, the numa_node of
pci devices under /sys/devices/pci0000:80/* is always 0, even if that
chain is on node 1 or 2 or 3.
Workaround: pcibus_to_node(bus) is used when we want to get the node that
pci_device is on.
In struct device, we already have numa_node member, and we could use
dev_to_node()/set_dev_node() to get and set numa_node in the device.
set_dev_node is called in pci_device_add() with pcibus_to_node(bus),
and pcibus_to_node uses bus->sysdata for nodeid.
The problem is when pci_add_device is called, bus->sysdata is not assigned
correct nodeid yet. The result is that numa_node will always be 0.
pcibios_scan_root and pci_scan_root could take sysdata. So we need to get
mp_bus_to_node mapping before these two are called, and thus
get_mp_bus_to_node could get correct node for sysdata in root bus.
In scanning of the root bus, all child busses will take parent bus sysdata.
So all pci_device->dev.numa_node will be assigned correctly and automatically.
Later we could use dev_to_node(&pci_dev->dev) to get numa_node, and we
could also could make other bus specific device get the correct numa_node
too.
This is an updated version of pci_sysdata and Jeff's pci_domain patch.
[ mingo@elte.hu: build fix ]
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-19 19:20:09 +08:00
|
|
|
unsigned slot;
|
2008-02-19 19:15:08 +08:00
|
|
|
int node;
|
2008-02-19 19:21:20 +08:00
|
|
|
int link;
|
|
|
|
int def_node;
|
|
|
|
int def_link;
|
|
|
|
struct pci_root_info *info;
|
|
|
|
u32 reg;
|
2010-02-10 17:20:10 +08:00
|
|
|
u64 start;
|
|
|
|
u64 end;
|
2010-02-10 17:20:07 +08:00
|
|
|
struct range range[RANGE_NUM];
|
2008-02-19 19:21:20 +08:00
|
|
|
u64 val;
|
|
|
|
u32 address;
|
2010-02-10 17:20:09 +08:00
|
|
|
bool found;
|
2012-01-06 05:27:19 +08:00
|
|
|
struct resource fam10h_mmconf_res, *fam10h_mmconf;
|
|
|
|
u64 fam10h_mmconf_start;
|
|
|
|
u64 fam10h_mmconf_end;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
x86: get mp_bus_to_node early
Currently, on an amd k8 system with multi ht chains, the numa_node of
pci devices under /sys/devices/pci0000:80/* is always 0, even if that
chain is on node 1 or 2 or 3.
Workaround: pcibus_to_node(bus) is used when we want to get the node that
pci_device is on.
In struct device, we already have numa_node member, and we could use
dev_to_node()/set_dev_node() to get and set numa_node in the device.
set_dev_node is called in pci_device_add() with pcibus_to_node(bus),
and pcibus_to_node uses bus->sysdata for nodeid.
The problem is when pci_add_device is called, bus->sysdata is not assigned
correct nodeid yet. The result is that numa_node will always be 0.
pcibios_scan_root and pci_scan_root could take sysdata. So we need to get
mp_bus_to_node mapping before these two are called, and thus
get_mp_bus_to_node could get correct node for sysdata in root bus.
In scanning of the root bus, all child busses will take parent bus sysdata.
So all pci_device->dev.numa_node will be assigned correctly and automatically.
Later we could use dev_to_node(&pci_dev->dev) to get numa_node, and we
could also could make other bus specific device get the correct numa_node
too.
This is an updated version of pci_sysdata and Jeff's pci_domain patch.
[ mingo@elte.hu: build fix ]
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-19 19:20:09 +08:00
|
|
|
if (!early_pci_allowed())
|
|
|
|
return -1;
|
|
|
|
|
2010-02-10 17:20:09 +08:00
|
|
|
found = false;
|
2008-02-19 19:21:20 +08:00
|
|
|
for (i = 0; i < ARRAY_SIZE(pci_probes); i++) {
|
|
|
|
u32 id;
|
|
|
|
u16 device;
|
|
|
|
u16 vendor;
|
2008-02-19 19:15:08 +08:00
|
|
|
|
2008-02-19 19:21:20 +08:00
|
|
|
bus = pci_probes[i].bus;
|
|
|
|
slot = pci_probes[i].slot;
|
|
|
|
id = read_pci_config(bus, slot, 0, PCI_VENDOR_ID);
|
2008-02-19 19:15:08 +08:00
|
|
|
|
2008-02-19 19:21:20 +08:00
|
|
|
vendor = id & 0xffff;
|
|
|
|
device = (id>>16) & 0xffff;
|
|
|
|
if (pci_probes[i].vendor == vendor &&
|
|
|
|
pci_probes[i].device == device) {
|
2010-02-10 17:20:09 +08:00
|
|
|
found = true;
|
2008-02-19 19:21:20 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-02-10 17:20:09 +08:00
|
|
|
if (!found)
|
2008-02-19 19:21:20 +08:00
|
|
|
return 0;
|
2008-02-19 19:15:08 +08:00
|
|
|
|
2008-02-19 19:21:20 +08:00
|
|
|
for (i = 0; i < 4; i++) {
|
|
|
|
int min_bus;
|
|
|
|
int max_bus;
|
|
|
|
reg = read_pci_config(bus, slot, 1, 0xe0 + (i << 2));
|
2008-02-19 19:15:08 +08:00
|
|
|
|
|
|
|
/* Check if that register is enabled for bus range */
|
2008-02-19 19:21:20 +08:00
|
|
|
if ((reg & 7) != 3)
|
2008-02-19 19:15:08 +08:00
|
|
|
continue;
|
|
|
|
|
2008-02-19 19:21:20 +08:00
|
|
|
min_bus = (reg >> 16) & 0xff;
|
|
|
|
max_bus = (reg >> 24) & 0xff;
|
|
|
|
node = (reg >> 4) & 0x07;
|
2012-05-02 07:25:18 +08:00
|
|
|
set_mp_bus_range_to_node(min_bus, max_bus, node);
|
2008-02-19 19:21:20 +08:00
|
|
|
link = (reg >> 8) & 0x03;
|
|
|
|
|
2012-04-03 09:31:54 +08:00
|
|
|
info = alloc_pci_root_info(min_bus, max_bus, node, link);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2008-02-19 19:21:20 +08:00
|
|
|
/* get the default node and link for left over res */
|
|
|
|
reg = read_pci_config(bus, slot, 0, 0x60);
|
|
|
|
def_node = (reg >> 8) & 0x07;
|
|
|
|
reg = read_pci_config(bus, slot, 0, 0x64);
|
|
|
|
def_link = (reg >> 8) & 0x03;
|
|
|
|
|
|
|
|
memset(range, 0, sizeof(range));
|
2010-02-10 17:20:13 +08:00
|
|
|
add_range(range, RANGE_NUM, 0, 0, 0xffff + 1);
|
2008-02-19 19:21:20 +08:00
|
|
|
/* io port resource */
|
|
|
|
for (i = 0; i < 4; i++) {
|
|
|
|
reg = read_pci_config(bus, slot, 1, 0xc0 + (i << 3));
|
|
|
|
if (!(reg & 3))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
start = reg & 0xfff000;
|
|
|
|
reg = read_pci_config(bus, slot, 1, 0xc4 + (i << 3));
|
|
|
|
node = reg & 0x07;
|
|
|
|
link = (reg >> 4) & 0x03;
|
|
|
|
end = (reg & 0xfff000) | 0xfff;
|
|
|
|
|
2012-04-03 09:31:54 +08:00
|
|
|
info = find_pci_root_info(node, link);
|
|
|
|
if (!info)
|
2008-02-19 19:21:20 +08:00
|
|
|
continue; /* not found */
|
|
|
|
|
2008-03-06 17:15:31 +08:00
|
|
|
printk(KERN_DEBUG "node %d link %d: io port [%llx, %llx]\n",
|
2010-02-10 17:20:10 +08:00
|
|
|
node, link, start, end);
|
2008-04-13 16:41:58 +08:00
|
|
|
|
|
|
|
/* kernel only handle 16 bit only */
|
|
|
|
if (end > 0xffff)
|
|
|
|
end = 0xffff;
|
|
|
|
update_res(info, start, end, IORESOURCE_IO, 1);
|
2010-02-10 17:20:13 +08:00
|
|
|
subtract_range(range, RANGE_NUM, start, end + 1);
|
2008-02-19 19:21:20 +08:00
|
|
|
}
|
|
|
|
/* add left over io port range to def node/link, [0, 0xffff] */
|
|
|
|
/* find the position */
|
2012-04-03 09:31:54 +08:00
|
|
|
info = find_pci_root_info(def_node, def_link);
|
|
|
|
if (info) {
|
2008-02-19 19:21:20 +08:00
|
|
|
for (i = 0; i < RANGE_NUM; i++) {
|
|
|
|
if (!range[i].end)
|
|
|
|
continue;
|
|
|
|
|
2010-02-10 17:20:13 +08:00
|
|
|
update_res(info, range[i].start, range[i].end - 1,
|
2008-02-19 19:21:20 +08:00
|
|
|
IORESOURCE_IO, 1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
memset(range, 0, sizeof(range));
|
|
|
|
/* 0xfd00000000-0xffffffffff for HT */
|
2010-02-10 17:20:13 +08:00
|
|
|
end = cap_resource((0xfdULL<<32) - 1);
|
|
|
|
end++;
|
|
|
|
add_range(range, RANGE_NUM, 0, 0, end);
|
2008-02-19 19:21:20 +08:00
|
|
|
|
|
|
|
/* need to take out [0, TOM) for RAM*/
|
|
|
|
address = MSR_K8_TOP_MEM1;
|
|
|
|
rdmsrl(address, val);
|
2008-05-13 08:40:39 +08:00
|
|
|
end = (val & 0xffffff800000ULL);
|
2010-02-10 17:20:10 +08:00
|
|
|
printk(KERN_INFO "TOM: %016llx aka %lldM\n", end, end>>20);
|
2008-02-19 19:21:20 +08:00
|
|
|
if (end < (1ULL<<32))
|
2010-02-10 17:20:13 +08:00
|
|
|
subtract_range(range, RANGE_NUM, 0, end);
|
2008-02-19 19:21:20 +08:00
|
|
|
|
2008-03-06 17:15:31 +08:00
|
|
|
/* get mmconfig */
|
2012-01-06 05:27:19 +08:00
|
|
|
fam10h_mmconf = amd_get_mmconfig_range(&fam10h_mmconf_res);
|
2008-03-06 17:15:31 +08:00
|
|
|
/* need to take out mmconf range */
|
2012-01-06 05:27:19 +08:00
|
|
|
if (fam10h_mmconf) {
|
|
|
|
printk(KERN_DEBUG "Fam 10h mmconf %pR\n", fam10h_mmconf);
|
|
|
|
fam10h_mmconf_start = fam10h_mmconf->start;
|
|
|
|
fam10h_mmconf_end = fam10h_mmconf->end;
|
2010-02-10 17:20:13 +08:00
|
|
|
subtract_range(range, RANGE_NUM, fam10h_mmconf_start,
|
|
|
|
fam10h_mmconf_end + 1);
|
2012-01-06 05:27:19 +08:00
|
|
|
} else {
|
|
|
|
fam10h_mmconf_start = 0;
|
|
|
|
fam10h_mmconf_end = 0;
|
2008-03-06 17:15:31 +08:00
|
|
|
}
|
|
|
|
|
2008-02-19 19:21:20 +08:00
|
|
|
/* mmio resource */
|
|
|
|
for (i = 0; i < 8; i++) {
|
|
|
|
reg = read_pci_config(bus, slot, 1, 0x80 + (i << 3));
|
|
|
|
if (!(reg & 3))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
start = reg & 0xffffff00; /* 39:16 on 31:8*/
|
|
|
|
start <<= 8;
|
|
|
|
reg = read_pci_config(bus, slot, 1, 0x84 + (i << 3));
|
|
|
|
node = reg & 0x07;
|
|
|
|
link = (reg >> 4) & 0x03;
|
|
|
|
end = (reg & 0xffffff00);
|
|
|
|
end <<= 8;
|
|
|
|
end |= 0xffff;
|
|
|
|
|
2012-04-03 09:31:54 +08:00
|
|
|
info = find_pci_root_info(node, link);
|
2008-02-19 19:21:20 +08:00
|
|
|
|
2012-04-03 09:31:54 +08:00
|
|
|
if (!info)
|
|
|
|
continue;
|
2008-03-06 17:15:31 +08:00
|
|
|
|
|
|
|
printk(KERN_DEBUG "node %d link %d: mmio [%llx, %llx]",
|
2010-02-10 17:20:10 +08:00
|
|
|
node, link, start, end);
|
2008-03-06 17:15:31 +08:00
|
|
|
/*
|
|
|
|
* some sick allocation would have range overlap with fam10h
|
|
|
|
* mmconf range, so need to update start and end.
|
|
|
|
*/
|
|
|
|
if (fam10h_mmconf_end) {
|
|
|
|
int changed = 0;
|
|
|
|
u64 endx = 0;
|
|
|
|
if (start >= fam10h_mmconf_start &&
|
|
|
|
start <= fam10h_mmconf_end) {
|
|
|
|
start = fam10h_mmconf_end + 1;
|
|
|
|
changed = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (end >= fam10h_mmconf_start &&
|
|
|
|
end <= fam10h_mmconf_end) {
|
|
|
|
end = fam10h_mmconf_start - 1;
|
|
|
|
changed = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (start < fam10h_mmconf_start &&
|
|
|
|
end > fam10h_mmconf_end) {
|
|
|
|
/* we got a hole */
|
|
|
|
endx = fam10h_mmconf_start - 1;
|
|
|
|
update_res(info, start, endx, IORESOURCE_MEM, 0);
|
2010-02-10 17:20:13 +08:00
|
|
|
subtract_range(range, RANGE_NUM, start,
|
|
|
|
endx + 1);
|
2010-02-10 17:20:10 +08:00
|
|
|
printk(KERN_CONT " ==> [%llx, %llx]", start, endx);
|
2008-03-06 17:15:31 +08:00
|
|
|
start = fam10h_mmconf_end + 1;
|
|
|
|
changed = 1;
|
|
|
|
}
|
|
|
|
if (changed) {
|
|
|
|
if (start <= end) {
|
2010-02-10 17:20:10 +08:00
|
|
|
printk(KERN_CONT " %s [%llx, %llx]", endx ? "and" : "==>", start, end);
|
2008-03-06 17:15:31 +08:00
|
|
|
} else {
|
|
|
|
printk(KERN_CONT "%s\n", endx?"":" ==> none");
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-02-10 17:20:11 +08:00
|
|
|
update_res(info, cap_resource(start), cap_resource(end),
|
|
|
|
IORESOURCE_MEM, 1);
|
2010-02-10 17:20:13 +08:00
|
|
|
subtract_range(range, RANGE_NUM, start, end + 1);
|
2008-03-06 17:15:31 +08:00
|
|
|
printk(KERN_CONT "\n");
|
2008-02-19 19:21:20 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/* need to take out [4G, TOM2) for RAM*/
|
|
|
|
/* SYS_CFG */
|
|
|
|
address = MSR_K8_SYSCFG;
|
|
|
|
rdmsrl(address, val);
|
|
|
|
/* TOP_MEM2 is enabled? */
|
|
|
|
if (val & (1<<21)) {
|
|
|
|
/* TOP_MEM2 */
|
|
|
|
address = MSR_K8_TOP_MEM2;
|
|
|
|
rdmsrl(address, val);
|
2008-05-13 08:40:39 +08:00
|
|
|
end = (val & 0xffffff800000ULL);
|
2010-02-10 17:20:10 +08:00
|
|
|
printk(KERN_INFO "TOM2: %016llx aka %lldM\n", end, end>>20);
|
2010-02-10 17:20:13 +08:00
|
|
|
subtract_range(range, RANGE_NUM, 1ULL<<32, end);
|
2008-02-19 19:21:20 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* add left over mmio range to def node/link ?
|
|
|
|
* that is tricky, just record range in from start_min to 4G
|
|
|
|
*/
|
2012-04-03 09:31:54 +08:00
|
|
|
info = find_pci_root_info(def_node, def_link);
|
|
|
|
if (info) {
|
2008-02-19 19:21:20 +08:00
|
|
|
for (i = 0; i < RANGE_NUM; i++) {
|
|
|
|
if (!range[i].end)
|
|
|
|
continue;
|
|
|
|
|
2010-02-10 17:20:11 +08:00
|
|
|
update_res(info, cap_resource(range[i].start),
|
2010-02-10 17:20:13 +08:00
|
|
|
cap_resource(range[i].end - 1),
|
2008-02-19 19:21:20 +08:00
|
|
|
IORESOURCE_MEM, 1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-04-03 09:31:54 +08:00
|
|
|
list_for_each_entry(info, &pci_root_infos, list) {
|
2008-02-19 19:21:20 +08:00
|
|
|
int busnum;
|
2012-04-03 09:31:54 +08:00
|
|
|
struct pci_root_res *root_res;
|
2008-02-19 19:21:20 +08:00
|
|
|
|
2012-05-18 09:51:12 +08:00
|
|
|
busnum = info->busn.start;
|
|
|
|
printk(KERN_DEBUG "bus: %pR on node %x link %x\n",
|
|
|
|
&info->busn, info->node, info->link);
|
2012-04-03 09:31:54 +08:00
|
|
|
list_for_each_entry(root_res, &info->resources, list)
|
|
|
|
printk(KERN_DEBUG "bus: %02x %pR\n",
|
|
|
|
busnum, &root_res->res);
|
2008-02-19 19:21:20 +08:00
|
|
|
}
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-06-13 02:19:23 +08:00
|
|
|
#define ENABLE_CF8_EXT_CFG (1ULL << 46)
|
|
|
|
|
x86: delete __cpuinit usage from all x86 files
The __cpuinit type of throwaway sections might have made sense
some time ago when RAM was more constrained, but now the savings
do not offset the cost and complications. For example, the fix in
commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time")
is a good example of the nasty type of bugs that can be created
with improper use of the various __init prefixes.
After a discussion on LKML[1] it was decided that cpuinit should go
the way of devinit and be phased out. Once all the users are gone,
we can then finally remove the macros themselves from linux/init.h.
Note that some harmless section mismatch warnings may result, since
notify_cpu_starting() and cpu_up() are arch independent (kernel/cpu.c)
are flagged as __cpuinit -- so if we remove the __cpuinit from
arch specific callers, we will also get section mismatch warnings.
As an intermediate step, we intend to turn the linux/init.h cpuinit
content into no-ops as early as possible, since that will get rid
of these warnings. In any case, they are temporary and harmless.
This removes all the arch/x86 uses of the __cpuinit macros from
all C files. x86 only had the one __CPUINIT used in assembly files,
and it wasn't paired off with a .previous or a __FINIT, so we can
delete it directly w/o any corresponding additional change there.
[1] https://lkml.org/lkml/2013/5/20/589
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: x86@kernel.org
Acked-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: H. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
2013-06-19 06:23:59 +08:00
|
|
|
static void enable_pci_io_ecs(void *unused)
|
2008-06-13 02:19:23 +08:00
|
|
|
{
|
|
|
|
u64 reg;
|
|
|
|
rdmsrl(MSR_AMD64_NB_CFG, reg);
|
|
|
|
if (!(reg & ENABLE_CF8_EXT_CFG)) {
|
|
|
|
reg |= ENABLE_CF8_EXT_CFG;
|
|
|
|
wrmsrl(MSR_AMD64_NB_CFG, reg);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
x86: delete __cpuinit usage from all x86 files
The __cpuinit type of throwaway sections might have made sense
some time ago when RAM was more constrained, but now the savings
do not offset the cost and complications. For example, the fix in
commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time")
is a good example of the nasty type of bugs that can be created
with improper use of the various __init prefixes.
After a discussion on LKML[1] it was decided that cpuinit should go
the way of devinit and be phased out. Once all the users are gone,
we can then finally remove the macros themselves from linux/init.h.
Note that some harmless section mismatch warnings may result, since
notify_cpu_starting() and cpu_up() are arch independent (kernel/cpu.c)
are flagged as __cpuinit -- so if we remove the __cpuinit from
arch specific callers, we will also get section mismatch warnings.
As an intermediate step, we intend to turn the linux/init.h cpuinit
content into no-ops as early as possible, since that will get rid
of these warnings. In any case, they are temporary and harmless.
This removes all the arch/x86 uses of the __cpuinit macros from
all C files. x86 only had the one __CPUINIT used in assembly files,
and it wasn't paired off with a .previous or a __FINIT, so we can
delete it directly w/o any corresponding additional change there.
[1] https://lkml.org/lkml/2013/5/20/589
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: x86@kernel.org
Acked-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: H. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
2013-06-19 06:23:59 +08:00
|
|
|
static int amd_cpu_notify(struct notifier_block *self, unsigned long action,
|
|
|
|
void *hcpu)
|
2008-06-13 02:19:23 +08:00
|
|
|
{
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
int cpu = (long)hcpu;
|
2008-08-23 02:23:38 +08:00
|
|
|
switch (action) {
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
case CPU_ONLINE:
|
|
|
|
case CPU_ONLINE_FROZEN:
|
|
|
|
smp_call_function_single(cpu, enable_pci_io_ecs, NULL, 0);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return NOTIFY_OK;
|
|
|
|
}
|
|
|
|
|
x86: delete __cpuinit usage from all x86 files
The __cpuinit type of throwaway sections might have made sense
some time ago when RAM was more constrained, but now the savings
do not offset the cost and complications. For example, the fix in
commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time")
is a good example of the nasty type of bugs that can be created
with improper use of the various __init prefixes.
After a discussion on LKML[1] it was decided that cpuinit should go
the way of devinit and be phased out. Once all the users are gone,
we can then finally remove the macros themselves from linux/init.h.
Note that some harmless section mismatch warnings may result, since
notify_cpu_starting() and cpu_up() are arch independent (kernel/cpu.c)
are flagged as __cpuinit -- so if we remove the __cpuinit from
arch specific callers, we will also get section mismatch warnings.
As an intermediate step, we intend to turn the linux/init.h cpuinit
content into no-ops as early as possible, since that will get rid
of these warnings. In any case, they are temporary and harmless.
This removes all the arch/x86 uses of the __cpuinit macros from
all C files. x86 only had the one __CPUINIT used in assembly files,
and it wasn't paired off with a .previous or a __FINIT, so we can
delete it directly w/o any corresponding additional change there.
[1] https://lkml.org/lkml/2013/5/20/589
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: x86@kernel.org
Acked-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: H. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
2013-06-19 06:23:59 +08:00
|
|
|
static struct notifier_block amd_cpu_notifier = {
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
.notifier_call = amd_cpu_notify,
|
|
|
|
};
|
|
|
|
|
2011-01-11 00:20:23 +08:00
|
|
|
static void __init pci_enable_pci_io_ecs(void)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_AMD_NB
|
|
|
|
unsigned int i, n;
|
|
|
|
|
|
|
|
for (n = i = 0; !n && amd_nb_bus_dev_ranges[i].dev_limit; ++i) {
|
|
|
|
u8 bus = amd_nb_bus_dev_ranges[i].bus;
|
|
|
|
u8 slot = amd_nb_bus_dev_ranges[i].dev_base;
|
|
|
|
u8 limit = amd_nb_bus_dev_ranges[i].dev_limit;
|
|
|
|
|
|
|
|
for (; slot < limit; ++slot) {
|
|
|
|
u32 val = read_pci_config(bus, slot, 3, 0);
|
|
|
|
|
|
|
|
if (!early_is_amd_nb(val))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
val = read_pci_config(bus, slot, 3, 0x8c);
|
|
|
|
if (!(val & (ENABLE_CF8_EXT_CFG >> 32))) {
|
|
|
|
val |= ENABLE_CF8_EXT_CFG >> 32;
|
|
|
|
write_pci_config(bus, slot, 3, 0x8c, val);
|
|
|
|
}
|
|
|
|
++n;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
static int __init pci_io_ecs_init(void)
|
|
|
|
{
|
|
|
|
int cpu;
|
|
|
|
|
2008-06-13 02:19:23 +08:00
|
|
|
/* assume all cpus from fam10h have IO ECS */
|
|
|
|
if (boot_cpu_data.x86 < 0x10)
|
|
|
|
return 0;
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
|
2011-01-11 00:20:23 +08:00
|
|
|
/* Try the PCI method first. */
|
|
|
|
if (early_pci_allowed())
|
|
|
|
pci_enable_pci_io_ecs();
|
|
|
|
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
register_cpu_notifier(&amd_cpu_notifier);
|
|
|
|
for_each_online_cpu(cpu)
|
|
|
|
amd_cpu_notify(&amd_cpu_notifier, (unsigned long)CPU_ONLINE,
|
|
|
|
(void *)(long)cpu);
|
2008-06-13 02:19:23 +08:00
|
|
|
pci_probe |= PCI_HAS_IO_ECS;
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
|
2008-06-13 02:19:23 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-08-23 02:23:37 +08:00
|
|
|
static int __init amd_postcore_init(void)
|
|
|
|
{
|
|
|
|
if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
early_fill_mp_bus_info();
|
x86: fix: make PCI ECS for AMD CPUs hotplug capable
Until now, PCI ECS setup was performed at boot time only and for cpus
that are enabled then. This patch fixes this and adds cpu hotplug.
Tests sequence (check if ECS bit is set when bringing cpu online again):
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
# ( perl -e 'sysseek(STDOUT, 0xC001001F, 0); print pack "l*", 8, 0x00400010' ) > /dev/cpu/1/msr
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00400010
# echo 0 > /sys/devices/system/cpu/cpu1/online
# echo 1 > /sys/devices/system/cpu/cpu1/online
# ( perl -e 'sysseek(STDIN, 0xC001001F, 0)'; hexdump -n 8 -e '2/4 "%08x " "\n"' ) < /dev/cpu/1/msr
00000008 00404010
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-23 02:23:38 +08:00
|
|
|
pci_io_ecs_init();
|
2008-08-23 02:23:37 +08:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
postcore_initcall(amd_postcore_init);
|