linux_dsm_epyc7002/arch/x86/mm/amdtopology.c

198 lines
4.3 KiB
C
Raw Normal View History

/*
* AMD NUMA support.
* Discover the memory map and associated nodes.
*
* This version reads it directly from the AMD northbridge.
*
* Copyright 2002,2003 Andi Kleen, SuSE Labs.
*/
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/module.h>
#include <linux/nodemask.h>
#include <linux/memblock.h>
#include <linux/bootmem.h>
#include <asm/io.h>
#include <linux/pci_ids.h>
#include <linux/acpi.h>
#include <asm/types.h>
#include <asm/mmzone.h>
#include <asm/proto.h>
#include <asm/e820.h>
#include <asm/pci-direct.h>
#include <asm/numa.h>
#include <asm/mpspec.h>
#include <asm/apic.h>
#include <asm/amd_nb.h>
static unsigned char __initdata nodeids[8];
x86: Export k8 physical topology To eventually interleave emulated nodes over physical nodes, we need to know the physical topology of the machine without actually registering it. This does the k8 node setup in two parts: detection and registration. NUMA emulation can then used the physical topology detected to setup the address ranges of emulated nodes accordingly. If emulation isn't used, the k8 nodes are registered as normal. Two formals are added to the x86 NUMA setup functions: `acpi' and `k8'. These represent whether ACPI or K8 NUMA has been detected; both cannot be true at the same time. This specifies to the NUMA emulation code whether an underlying physical NUMA topology exists and which interface to use. This patch deals solely with separating the k8 setup path into Northbridge detection and registration steps and leaves the ACPI changes for a subsequent patch. The `acpi' formal is added here, however, to avoid touching all the header files again in the next patch. This approach also ensures emulated nodes will not span physical nodes so the true memory latency is not misrepresented. k8_get_nodes() may now be used to export the k8 physical topology of the machine for NUMA emulation. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: Yinghai Lu <yinghai@kernel.org> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Ankita Garg <ankita@in.ibm.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <alpine.DEB.1.00.0909251518400.14754@chino.kir.corp.google.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-26 05:20:00 +07:00
static __init int find_northbridge(void)
{
int num;
for (num = 0; num < 32; num++) {
u32 header;
header = read_pci_config(0, num, 0, 0x00);
if (header != (PCI_VENDOR_ID_AMD | (0x1100<<16)) &&
header != (PCI_VENDOR_ID_AMD | (0x1200<<16)) &&
header != (PCI_VENDOR_ID_AMD | (0x1300<<16)))
continue;
header = read_pci_config(0, num, 1, 0x00);
if (header != (PCI_VENDOR_ID_AMD | (0x1101<<16)) &&
header != (PCI_VENDOR_ID_AMD | (0x1201<<16)) &&
header != (PCI_VENDOR_ID_AMD | (0x1301<<16)))
continue;
return num;
}
return -ENOENT;
}
static __init void early_get_boot_cpu_id(void)
{
/*
* need to get the APIC ID of the BSP so can use that to
* create apicid_to_node in amd_scan_nodes()
*/
#ifdef CONFIG_X86_MPPARSE
/*
* get boot-time SMP configuration:
*/
if (smp_found_config)
early_get_smp_config();
#endif
x86: Export k8 physical topology To eventually interleave emulated nodes over physical nodes, we need to know the physical topology of the machine without actually registering it. This does the k8 node setup in two parts: detection and registration. NUMA emulation can then used the physical topology detected to setup the address ranges of emulated nodes accordingly. If emulation isn't used, the k8 nodes are registered as normal. Two formals are added to the x86 NUMA setup functions: `acpi' and `k8'. These represent whether ACPI or K8 NUMA has been detected; both cannot be true at the same time. This specifies to the NUMA emulation code whether an underlying physical NUMA topology exists and which interface to use. This patch deals solely with separating the k8 setup path into Northbridge detection and registration steps and leaves the ACPI changes for a subsequent patch. The `acpi' formal is added here, however, to avoid touching all the header files again in the next patch. This approach also ensures emulated nodes will not span physical nodes so the true memory latency is not misrepresented. k8_get_nodes() may now be used to export the k8 physical topology of the machine for NUMA emulation. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: Yinghai Lu <yinghai@kernel.org> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Ankita Garg <ankita@in.ibm.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <alpine.DEB.1.00.0909251518400.14754@chino.kir.corp.google.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-26 05:20:00 +07:00
}
int __init amd_numa_init(void)
x86: Export k8 physical topology To eventually interleave emulated nodes over physical nodes, we need to know the physical topology of the machine without actually registering it. This does the k8 node setup in two parts: detection and registration. NUMA emulation can then used the physical topology detected to setup the address ranges of emulated nodes accordingly. If emulation isn't used, the k8 nodes are registered as normal. Two formals are added to the x86 NUMA setup functions: `acpi' and `k8'. These represent whether ACPI or K8 NUMA has been detected; both cannot be true at the same time. This specifies to the NUMA emulation code whether an underlying physical NUMA topology exists and which interface to use. This patch deals solely with separating the k8 setup path into Northbridge detection and registration steps and leaves the ACPI changes for a subsequent patch. The `acpi' formal is added here, however, to avoid touching all the header files again in the next patch. This approach also ensures emulated nodes will not span physical nodes so the true memory latency is not misrepresented. k8_get_nodes() may now be used to export the k8 physical topology of the machine for NUMA emulation. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: Yinghai Lu <yinghai@kernel.org> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Ankita Garg <ankita@in.ibm.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <alpine.DEB.1.00.0909251518400.14754@chino.kir.corp.google.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-26 05:20:00 +07:00
{
u64 start = PFN_PHYS(0);
u64 end = PFN_PHYS(max_pfn);
x86: Export k8 physical topology To eventually interleave emulated nodes over physical nodes, we need to know the physical topology of the machine without actually registering it. This does the k8 node setup in two parts: detection and registration. NUMA emulation can then used the physical topology detected to setup the address ranges of emulated nodes accordingly. If emulation isn't used, the k8 nodes are registered as normal. Two formals are added to the x86 NUMA setup functions: `acpi' and `k8'. These represent whether ACPI or K8 NUMA has been detected; both cannot be true at the same time. This specifies to the NUMA emulation code whether an underlying physical NUMA topology exists and which interface to use. This patch deals solely with separating the k8 setup path into Northbridge detection and registration steps and leaves the ACPI changes for a subsequent patch. The `acpi' formal is added here, however, to avoid touching all the header files again in the next patch. This approach also ensures emulated nodes will not span physical nodes so the true memory latency is not misrepresented. k8_get_nodes() may now be used to export the k8 physical topology of the machine for NUMA emulation. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: Yinghai Lu <yinghai@kernel.org> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Ankita Garg <ankita@in.ibm.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <alpine.DEB.1.00.0909251518400.14754@chino.kir.corp.google.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-26 05:20:00 +07:00
unsigned numnodes;
u64 prevbase;
int i, j, nb;
u32 nodeid, reg;
unsigned int bits, cores, apicid_base;
if (!early_pci_allowed())
return -EINVAL;
nb = find_northbridge();
if (nb < 0)
return nb;
pr_info("Scanning NUMA topology in Northbridge %d\n", nb);
reg = read_pci_config(0, nb, 0, 0x60);
numnodes = ((reg >> 4) & 0xF) + 1;
if (numnodes <= 1)
return -ENOENT;
x86: Export k8 physical topology To eventually interleave emulated nodes over physical nodes, we need to know the physical topology of the machine without actually registering it. This does the k8 node setup in two parts: detection and registration. NUMA emulation can then used the physical topology detected to setup the address ranges of emulated nodes accordingly. If emulation isn't used, the k8 nodes are registered as normal. Two formals are added to the x86 NUMA setup functions: `acpi' and `k8'. These represent whether ACPI or K8 NUMA has been detected; both cannot be true at the same time. This specifies to the NUMA emulation code whether an underlying physical NUMA topology exists and which interface to use. This patch deals solely with separating the k8 setup path into Northbridge detection and registration steps and leaves the ACPI changes for a subsequent patch. The `acpi' formal is added here, however, to avoid touching all the header files again in the next patch. This approach also ensures emulated nodes will not span physical nodes so the true memory latency is not misrepresented. k8_get_nodes() may now be used to export the k8 physical topology of the machine for NUMA emulation. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: Yinghai Lu <yinghai@kernel.org> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Ankita Garg <ankita@in.ibm.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <alpine.DEB.1.00.0909251518400.14754@chino.kir.corp.google.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-26 05:20:00 +07:00
pr_info("Number of physical nodes %d\n", numnodes);
prevbase = 0;
for (i = 0; i < 8; i++) {
u64 base, limit;
base = read_pci_config(0, nb, 1, 0x40 + i*8);
limit = read_pci_config(0, nb, 1, 0x44 + i*8);
nodeids[i] = nodeid = limit & 7;
if ((base & 3) == 0) {
if (i < numnodes)
pr_info("Skipping disabled node %d\n", i);
continue;
}
if (nodeid >= numnodes) {
pr_info("Ignoring excess node %d (%Lx:%Lx)\n", nodeid,
base, limit);
continue;
}
if (!limit) {
pr_info("Skipping node entry %d (base %Lx)\n",
i, base);
continue;
}
if ((base >> 8) & 3 || (limit >> 8) & 3) {
pr_err("Node %d using interleaving mode %Lx/%Lx\n",
nodeid, (base >> 8) & 3, (limit >> 8) & 3);
return -EINVAL;
}
if (node_isset(nodeid, numa_nodes_parsed)) {
pr_info("Node %d already present, skipping\n",
nodeid);
continue;
}
limit >>= 16;
limit <<= 24;
limit |= (1<<24)-1;
limit++;
x86: Export k8 physical topology To eventually interleave emulated nodes over physical nodes, we need to know the physical topology of the machine without actually registering it. This does the k8 node setup in two parts: detection and registration. NUMA emulation can then used the physical topology detected to setup the address ranges of emulated nodes accordingly. If emulation isn't used, the k8 nodes are registered as normal. Two formals are added to the x86 NUMA setup functions: `acpi' and `k8'. These represent whether ACPI or K8 NUMA has been detected; both cannot be true at the same time. This specifies to the NUMA emulation code whether an underlying physical NUMA topology exists and which interface to use. This patch deals solely with separating the k8 setup path into Northbridge detection and registration steps and leaves the ACPI changes for a subsequent patch. The `acpi' formal is added here, however, to avoid touching all the header files again in the next patch. This approach also ensures emulated nodes will not span physical nodes so the true memory latency is not misrepresented. k8_get_nodes() may now be used to export the k8 physical topology of the machine for NUMA emulation. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: Yinghai Lu <yinghai@kernel.org> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Ankita Garg <ankita@in.ibm.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <alpine.DEB.1.00.0909251518400.14754@chino.kir.corp.google.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-26 05:20:00 +07:00
if (limit > end)
limit = end;
if (limit <= base)
continue;
base >>= 16;
base <<= 24;
if (base < start)
base = start;
if (limit > end)
limit = end;
if (limit == base) {
pr_err("Empty node %d\n", nodeid);
continue;
}
if (limit < base) {
pr_err("Node %d bogus settings %Lx-%Lx.\n",
nodeid, base, limit);
continue;
}
/* Could sort here, but pun for now. Should not happen anyroads. */
if (prevbase > base) {
pr_err("Node map not sorted %Lx,%Lx\n",
prevbase, base);
return -EINVAL;
}
pr_info("Node %d MemBase %016Lx Limit %016Lx\n",
nodeid, base, limit);
prevbase = base;
numa_add_memblk(nodeid, base, limit);
node_set(nodeid, numa_nodes_parsed);
}
if (!nodes_weight(numa_nodes_parsed))
return -ENOENT;
/*
* We seem to have valid NUMA configuration. Map apicids to nodes
* using the coreid bits from early_identify_cpu.
*/
bits = boot_cpu_data.x86_coreid_bits;
cores = 1 << bits;
apicid_base = 0;
/* get the APIC ID of the BSP early for systems with apicid lifting */
early_get_boot_cpu_id();
if (boot_cpu_physical_apicid > 0) {
pr_info("BSP APIC ID: %02x\n", boot_cpu_physical_apicid);
apicid_base = boot_cpu_physical_apicid;
}
for_each_node_mask(i, numa_nodes_parsed)
for (j = apicid_base; j < cores + apicid_base; j++)
set_apicid_to_node((i << bits) + j, i);
x86: Export k8 physical topology To eventually interleave emulated nodes over physical nodes, we need to know the physical topology of the machine without actually registering it. This does the k8 node setup in two parts: detection and registration. NUMA emulation can then used the physical topology detected to setup the address ranges of emulated nodes accordingly. If emulation isn't used, the k8 nodes are registered as normal. Two formals are added to the x86 NUMA setup functions: `acpi' and `k8'. These represent whether ACPI or K8 NUMA has been detected; both cannot be true at the same time. This specifies to the NUMA emulation code whether an underlying physical NUMA topology exists and which interface to use. This patch deals solely with separating the k8 setup path into Northbridge detection and registration steps and leaves the ACPI changes for a subsequent patch. The `acpi' formal is added here, however, to avoid touching all the header files again in the next patch. This approach also ensures emulated nodes will not span physical nodes so the true memory latency is not misrepresented. k8_get_nodes() may now be used to export the k8 physical topology of the machine for NUMA emulation. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: Yinghai Lu <yinghai@kernel.org> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Ankita Garg <ankita@in.ibm.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <alpine.DEB.1.00.0909251518400.14754@chino.kir.corp.google.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-26 05:20:00 +07:00
return 0;
}