x86, amd: Add basic support for AMD CPU family 15h
cover x86, amd: Add basic support for AMD CPU family 15h
Hi,
following a bunch of patches with basic adaptions to support AMD CPU family 15h.
Patches are against tip/master (v2.6.36-rc6-1544-gc5bf3de). Please apply.
Thanks,
Andreas
PATCH 1/6 x86, mtrr: Assume SYS_CFG[Tom2ForceMemTypeWB] exists on all future AMD CPUs
From: Andreas Herrmann <[email protected]>
Instead of adapting the CPU family check in amd_special_default_mtrr() for each new CPU family assume that all new AMD CPUs support the necessary bits in SYS_CFG MSR.
Tom2Enabled is architectural (defined in APM Vol.2). Tom2ForceMemTypeWB is defined in all BKDGs starting with K8 NPT. In pre K8-NPT BKDG this bit is reserved (read as zero).
W/o this adaption Linux would unnecessarily complain about bad MTRR settings on every new AMD CPU family, e.g.
[ 0.000000] WARNING: BIOS bug: CPU MTRRs don't cover all of memory, losing 4863MB of RAM.
--- arch/x86/kernel/cpu/mtrr/cleanup.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
arch/x86/kernel/cpu/mtrr/cleanup.c +1 -1
@@ -827,7 +827,7 @@ int __init amd_special_default_mtrr(void)if (boot_cpu_data.x86_vendor !=X86_VENDOR_AMD )return 0 ; -if (boot_cpu_data.x86 <0xf || boot_cpu_data.x86 >0x11 ) +if (boot_cpu_data.x86 <0xf )return 0 ;/* In case some hypervisor doesn't pass SYSCFG through: */ if (rdmsr_safe (MSR_K8_SYSCFG , &l, &h) <0 )
Signature · 1 line
1.6.4.4
[tip:x86/amd-nb] x86, mtrr: Assume SYS_CFG[Tom2ForceMemTypeWB] exists on all future AMD CPUs
Commit-ID: 3fdbf004c1706480a7c7fac3c9d836fa6df20d7d Gitweb: http://git.kernel.org/tip/3fdbf004c1706480a7c7fac3c9d836fa6df20d7d Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 30 Sep 2010 14:32:35 +0200 Committer: H. Peter Anvin <[email protected]> CommitDate: Fri, 1 Oct 2010 16:18:31 -0700
x86, mtrr: Assume SYS_CFG[Tom2ForceMemTypeWB] exists on all future AMD CPUs
Instead of adapting the CPU family check in amd_special_default_mtrr() for each new CPU family assume that all new AMD CPUs support the necessary bits in SYS_CFG MSR.
Tom2Enabled is architectural (defined in APM Vol.2). Tom2ForceMemTypeWB is defined in all BKDGs starting with K8 NPT. In pre K8-NPT BKDG this bit is reserved (read as zero).
W/o this adaption Linux would unnecessarily complain about bad MTRR settings on every new AMD CPU family, e.g.
[ 0.000000] WARNING: BIOS bug: CPU MTRRs don't cover all of memory, losing 4863MB of RAM.
LKML-Reference: <[email protected]>
--- arch/x86/kernel/cpu/mtrr/cleanup.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
arch/x86/kernel/cpu/mtrr/cleanup.c +1 -1
@@ -827,7 +827,7 @@ int __init amd_special_default_mtrr(void)if (boot_cpu_data.x86_vendor !=X86_VENDOR_AMD )return 0 ; -if (boot_cpu_data.x86 <0xf || boot_cpu_data.x86 >0x11 ) +if (boot_cpu_data.x86 <0xf )return 0 ;/* In case some hypervisor doesn't pass SYSCFG through: */ if (rdmsr_safe (MSR_K8_SYSCFG , &l, &h) <0 )
PATCH 2/6 x86, nmi: Support NMI watchdog on newer AMD CPU families
In reply to Andreas HerrmannFrom: Andreas Herrmann <[email protected]>
CPU families 0x12, 0x14 and 0x15 support this functionality.
--- arch/x86/kernel/cpu/perfctr-watchdog.c | 9 ++++----- 1 files changed, 4 insertions(+), 5 deletions(-)
arch/x86/kernel/cpu/perfctr-watchdog.c +4 -5
@@ -700,11 +700,10 @@ static void probe_nmi_watchdog(void) {switch (boot_cpu_data.x86_vendor) {case X86_VENDOR_AMD : -if (boot_cpu_data.x86 !=6 && boot_cpu_data.x86 !=15 && - boot_cpu_data.x86 !=16 && boot_cpu_data.x86 !=17 ) -return ; - wd_ops = &k7_wd_ops; -break ; +if (boot_cpu_data.x86 ==6 || + (boot_cpu_data.x86 >=0xf && boot_cpu_data.x86 <=0x15 )) + wd_ops = &k7_wd_ops; +return ;case X86_VENDOR_INTEL :/* Work around where perfctr1 doesn't have a working enable * bit as described in the following errata:
Signature · 1 line
1.6.4.4
[tip:x86/amd-nb] x86, nmi: Support NMI watchdog on newer AMD CPU families
Commit-ID: 420b13b60a3e5c5dcc6ec290e131cf5fbc603d94 Gitweb: http://git.kernel.org/tip/420b13b60a3e5c5dcc6ec290e131cf5fbc603d94 Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 30 Sep 2010 14:33:58 +0200 Committer: H. Peter Anvin <[email protected]> CommitDate: Fri, 1 Oct 2010 16:18:32 -0700
x86, nmi: Support NMI watchdog on newer AMD CPU families
CPU families 0x12, 0x14 and 0x15 support this functionality.
LKML-Reference: <[email protected]>
--- arch/x86/kernel/cpu/perfctr-watchdog.c | 9 ++++----- 1 files changed, 4 insertions(+), 5 deletions(-)
arch/x86/kernel/cpu/perfctr-watchdog.c +4 -5
@@ -700,11 +700,10 @@ static void probe_nmi_watchdog(void) {switch (boot_cpu_data.x86_vendor) {case X86_VENDOR_AMD : -if (boot_cpu_data.x86 !=6 && boot_cpu_data.x86 !=15 && - boot_cpu_data.x86 !=16 && boot_cpu_data.x86 !=17 ) -return ; - wd_ops = &k7_wd_ops; -break ; +if (boot_cpu_data.x86 ==6 || + (boot_cpu_data.x86 >=0xf && boot_cpu_data.x86 <=0x15 )) + wd_ops = &k7_wd_ops; +return ;case X86_VENDOR_INTEL :/* Work around where perfctr1 doesn't have a working enable * bit as described in the following errata:
PATCH 3/6 x86, amd: Add support for CPUID topology extension of AMD CPUs
In reply to Andreas HerrmannFrom: Andreas Herrmann <[email protected]>
Node information (ID, number of internal nodes) is provided via CPUID Fn8000_001e_ECX.
See AMD CPUID Specification (Publication # 25481, Revision 2.34, September 2010).
--- arch/x86/kernel/cpu/amd.c | 50 +++++++++++++++++++++++--------------------- 1 files changed, 26 insertions(+), 24 deletions(-)
arch/x86/kernel/cpu/amd.c +26 -24
@@ -253,37 +253,41 @@ static int __cpuinit nearby_node(int apicid)#endif /* -* Fixup core topology information for AMD multi-node processors. -* Assumption: Number of cores in each internal node is the same. +* Fixup core topology information for +* (1) AMD multi-node processors +* Assumption: Number of cores in each internal node is the same. */ #ifdef CONFIG_X86_HT -static void __cpuinit amd_fixup_dcm (struct cpuinfo_x86 *c) +static void __cpuinit amd_get_topology (struct cpuinfo_x86 *c) { -unsigned long long value;u32 nodes, cores_per_node; +u8 node_id; +unsigned long long value;int cpu =smp_processor_id (); -if (!cpu_has (c,X86_FEATURE_NODEID_MSR )) -return ; - -/* fixup topology information only once for a core */ -if (cpu_has (c,X86_FEATURE_AMD_DCM )) +/* get information required for multi-node processors */ +if (cpu_has (c,X86_FEATURE_TOPOEXT )) { + value =cpuid_ecx (0x8000001e ); + nodes = ((value >>8 ) &7 ) +1 ; + node_id = value &7 ; + }else if (cpu_has (c,X86_FEATURE_NODEID_MSR )) { +rdmsrl (MSR_FAM10H_NODE_ID , value); + nodes = ((value >>3 ) &7 ) +1 ; + node_id = value &7 ; + }else return ; -rdmsrl (MSR_FAM10H_NODE_ID , value); +/* fixup multi-node processor information */ +if (nodes >1 ) { +set_cpu_cap (c,X86_FEATURE_AMD_DCM ); + cores_per_node = c->x86_max_cores / nodes; - nodes = ((value >>3 ) &7 ) +1 ; -if (nodes ==1 ) -return ; - -set_cpu_cap (c,X86_FEATURE_AMD_DCM ); - cores_per_node = c->x86_max_cores / nodes; +/* store NodeID, use llc_shared_map to store sibling info */ +per_cpu (cpu_llc_id, cpu) = node_id; -/* store NodeID, use llc_shared_map to store sibling info */ -per_cpu (cpu_llc_id, cpu) = value &7 ; - -/* fixup core id to be in range from 0 to (cores_per_node - 1) */ - c->cpu_core_id = c->cpu_core_id % cores_per_node; +/* core id to be in range from 0 to (cores_per_node - 1) */ + c->cpu_core_id = c->cpu_core_id % cores_per_node; + } }#endif
@@ -304,9 +308,7 @@ static void __cpuinit amd_detect_cmp(struct cpuinfo_x86 *c) c->phys_proc_id = c->initial_apicid >> bits;/* use socket ID also for last level cache */ per_cpu (cpu_llc_id, cpu) = c->phys_proc_id; -/* fixup topology information on multi-node processors */ -if ((c->x86 ==0x10 ) && (c->x86_model ==9 )) -amd_fixup_dcm (c); +amd_get_topology (c);#endif }
Signature · 1 line
1.6.4.4
[tip:x86/amd-nb] x86, amd: Add support for CPUID topology extension of AMD CPUs
Commit-ID: 23588c38a84c9175c6668789b64ffba4651e5c6a Gitweb: http://git.kernel.org/tip/23588c38a84c9175c6668789b64ffba4651e5c6a Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 30 Sep 2010 14:36:28 +0200 Committer: H. Peter Anvin <[email protected]> CommitDate: Fri, 1 Oct 2010 16:18:32 -0700
x86, amd: Add support for CPUID topology extension of AMD CPUs
Node information (ID, number of internal nodes) is provided via CPUID Fn8000_001e_ECX.
See AMD CPUID Specification (Publication # 25481, Revision 2.34, September 2010).
LKML-Reference: <[email protected]>
--- arch/x86/kernel/cpu/amd.c | 50 +++++++++++++++++++++++--------------------- 1 files changed, 26 insertions(+), 24 deletions(-)
arch/x86/kernel/cpu/amd.c +26 -24
@@ -253,37 +253,41 @@ static int __cpuinit nearby_node(int apicid)#endif /* -* Fixup core topology information for AMD multi-node processors. -* Assumption: Number of cores in each internal node is the same. +* Fixup core topology information for +* (1) AMD multi-node processors +* Assumption: Number of cores in each internal node is the same. */ #ifdef CONFIG_X86_HT -static void __cpuinit amd_fixup_dcm (struct cpuinfo_x86 *c) +static void __cpuinit amd_get_topology (struct cpuinfo_x86 *c) { -unsigned long long value;u32 nodes, cores_per_node; +u8 node_id; +unsigned long long value;int cpu =smp_processor_id (); -if (!cpu_has (c,X86_FEATURE_NODEID_MSR )) -return ; - -/* fixup topology information only once for a core */ -if (cpu_has (c,X86_FEATURE_AMD_DCM )) +/* get information required for multi-node processors */ +if (cpu_has (c,X86_FEATURE_TOPOEXT )) { + value =cpuid_ecx (0x8000001e ); + nodes = ((value >>8 ) &7 ) +1 ; + node_id = value &7 ; + }else if (cpu_has (c,X86_FEATURE_NODEID_MSR )) { +rdmsrl (MSR_FAM10H_NODE_ID , value); + nodes = ((value >>3 ) &7 ) +1 ; + node_id = value &7 ; + }else return ; -rdmsrl (MSR_FAM10H_NODE_ID , value); +/* fixup multi-node processor information */ +if (nodes >1 ) { +set_cpu_cap (c,X86_FEATURE_AMD_DCM ); + cores_per_node = c->x86_max_cores / nodes; - nodes = ((value >>3 ) &7 ) +1 ; -if (nodes ==1 ) -return ; - -set_cpu_cap (c,X86_FEATURE_AMD_DCM ); - cores_per_node = c->x86_max_cores / nodes; +/* store NodeID, use llc_shared_map to store sibling info */ +per_cpu (cpu_llc_id, cpu) = node_id; -/* store NodeID, use llc_shared_map to store sibling info */ -per_cpu (cpu_llc_id, cpu) = value &7 ; - -/* fixup core id to be in range from 0 to (cores_per_node - 1) */ - c->cpu_core_id = c->cpu_core_id % cores_per_node; +/* core id to be in range from 0 to (cores_per_node - 1) */ + c->cpu_core_id = c->cpu_core_id % cores_per_node; + } }#endif
@@ -304,9 +308,7 @@ static void __cpuinit amd_detect_cmp(struct cpuinfo_x86 *c) c->phys_proc_id = c->initial_apicid >> bits;/* use socket ID also for last level cache */ per_cpu (cpu_llc_id, cpu) = c->phys_proc_id; -/* fixup topology information on multi-node processors */ -if ((c->x86 ==0x10 ) && (c->x86_model ==9 )) -amd_fixup_dcm (c); +amd_get_topology (c);#endif }
PATCH 4/6 x86, amd: Extract compute unit information for AMD CPUs
In reply to Andreas HerrmannFrom: Andreas Herrmann <[email protected]>
Get compute unit information from CPUID Fn8000_001E_EBX. (See AMD CPUID Specification - publication # 25481, revision 2.34, September 2010.)
Note that each core on a compute unit still has a core_id of its own.
--- arch/x86/include/asm/processor.h | 2 ++ arch/x86/kernel/cpu/amd.c | 20 +++++++++++++++----- 2 files changed, 17 insertions(+), 5 deletions(-)
arch/x86/include/asm/processor.h +2 -0
@@ -110,6 +110,8 @@ struct cpuinfo_x86 {u16 phys_proc_id;/* Core id: */ u16 cpu_core_id; +/* Compute unit id */ +u8 compute_unit_id;/* Index into per_cpu list: */ u16 cpu_index;#endif
arch/x86/kernel/cpu/amd.c +15 -5
@@ -256,21 +256,29 @@ static int __cpuinit nearby_node(int apicid)* Fixup core topology information for * (1) AMD multi-node processors * Assumption: Number of cores in each internal node is the same. +* (2) AMD processors supporting compute units */ #ifdef CONFIG_X86_HT static void __cpuinit amd_get_topology (struct cpuinfo_x86 *c) { -u32 nodes, cores_per_node; +u32 nodes;u8 node_id; -unsigned long long value;int cpu =smp_processor_id ();/* get information required for multi-node processors */ if (cpu_has (c,X86_FEATURE_TOPOEXT )) { - value =cpuid_ecx (0x8000001e ); - nodes = ((value >>8 ) &7 ) +1 ; - node_id = value &7 ; +u32 eax, ebx, ecx, edx; + +cpuid (0x8000001e , &eax, &ebx, &ecx, &edx); + nodes = ((ecx >>8 ) &7 ) +1 ; + node_id = ecx &7 ; + +/* get compute unit information */ + smp_num_siblings = ((ebx >>8 ) &3 ) +1 ; + c->compute_unit_id = ebx &0xff ; }else if (cpu_has (c,X86_FEATURE_NODEID_MSR )) { +u64 value; +rdmsrl (MSR_FAM10H_NODE_ID , value); nodes = ((value >>3 ) &7 ) +1 ; node_id = value &7 ;
@@ -279,6 +287,8 @@ static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c)/* fixup multi-node processor information */ if (nodes >1 ) { +u32 cores_per_node; +set_cpu_cap (c,X86_FEATURE_AMD_DCM ); cores_per_node = c->x86_max_cores / nodes;
Signature · 1 line
1.6.4.4
[tip:x86/amd-nb] x86, amd: Extract compute unit information for AMD CPUs
Commit-ID: 6057b4d331f19a3ea51aec463ea7839c128b3227 Gitweb: http://git.kernel.org/tip/6057b4d331f19a3ea51aec463ea7839c128b3227 Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 30 Sep 2010 14:38:57 +0200 Committer: H. Peter Anvin <[email protected]> CommitDate: Fri, 1 Oct 2010 16:18:32 -0700
x86, amd: Extract compute unit information for AMD CPUs
Get compute unit information from CPUID Fn8000_001E_EBX. (See AMD CPUID Specification - publication # 25481, revision 2.34, September 2010.)
Note that each core on a compute unit still has a core_id of its own.
LKML-Reference: <[email protected]>
--- arch/x86/include/asm/processor.h | 2 ++ arch/x86/kernel/cpu/amd.c | 20 +++++++++++++++----- 2 files changed, 17 insertions(+), 5 deletions(-)
arch/x86/include/asm/processor.h +2 -0
@@ -110,6 +110,8 @@ struct cpuinfo_x86 {u16 phys_proc_id;/* Core id: */ u16 cpu_core_id; +/* Compute unit id */ +u8 compute_unit_id;/* Index into per_cpu list: */ u16 cpu_index;#endif
arch/x86/kernel/cpu/amd.c +15 -5
@@ -256,21 +256,29 @@ static int __cpuinit nearby_node(int apicid)* Fixup core topology information for * (1) AMD multi-node processors * Assumption: Number of cores in each internal node is the same. +* (2) AMD processors supporting compute units */ #ifdef CONFIG_X86_HT static void __cpuinit amd_get_topology (struct cpuinfo_x86 *c) { -u32 nodes, cores_per_node; +u32 nodes;u8 node_id; -unsigned long long value;int cpu =smp_processor_id ();/* get information required for multi-node processors */ if (cpu_has (c,X86_FEATURE_TOPOEXT )) { - value =cpuid_ecx (0x8000001e ); - nodes = ((value >>8 ) &7 ) +1 ; - node_id = value &7 ; +u32 eax, ebx, ecx, edx; + +cpuid (0x8000001e , &eax, &ebx, &ecx, &edx); + nodes = ((ecx >>8 ) &7 ) +1 ; + node_id = ecx &7 ; + +/* get compute unit information */ + smp_num_siblings = ((ebx >>8 ) &3 ) +1 ; + c->compute_unit_id = ebx &0xff ; }else if (cpu_has (c,X86_FEATURE_NODEID_MSR )) { +u64 value; +rdmsrl (MSR_FAM10H_NODE_ID , value); nodes = ((value >>3 ) &7 ) +1 ; node_id = value &7 ;
@@ -279,6 +287,8 @@ static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c)/* fixup multi-node processor information */ if (nodes >1 ) { +u32 cores_per_node; +set_cpu_cap (c,X86_FEATURE_AMD_DCM ); cores_per_node = c->x86_max_cores / nodes;
PATCH 5/6 x86, amd: Use compute unit information to determine thread siblings
In reply to Andreas HerrmannFrom: Andreas Herrmann <[email protected]>
This information is vital for different load balancing policies.
--- arch/x86/kernel/smpboot.c | 28 ++++++++++++++++++++-------- 1 files changed, 20 insertions(+), 8 deletions(-)
arch/x86/kernel/smpboot.c +20 -8
@@ -396,6 +396,19 @@ void __cpuinit smp_store_cpu_info(int id)identify_secondary_cpu (c); } +static void __cpuinit link_thread_siblings (int cpu1,int cpu2) +{ +struct cpuinfo_x86 *c1 = &cpu_data (cpu1); +struct cpuinfo_x86 *c2 = &cpu_data (cpu2); + +cpumask_set_cpu (cpu1,cpu_sibling_mask (cpu2)); +cpumask_set_cpu (cpu2,cpu_sibling_mask (cpu1)); +cpumask_set_cpu (cpu1,cpu_core_mask (cpu2)); +cpumask_set_cpu (cpu2,cpu_core_mask (cpu1)); +cpumask_set_cpu (cpu1, c2->llc_shared_map); +cpumask_set_cpu (cpu2, c1->llc_shared_map); +} +void __cpuinit set_cpu_sibling_map (int cpu) {
@@ -408,14 +421,13 @@ void __cpuinit set_cpu_sibling_map(int cpu)for_each_cpu (i, cpu_sibling_setup_mask) {struct cpuinfo_x86 *o = &cpu_data (i); -if (c->phys_proc_id == o->phys_proc_id && - c->cpu_core_id == o->cpu_core_id) { -cpumask_set_cpu (i,cpu_sibling_mask (cpu)); -cpumask_set_cpu (cpu,cpu_sibling_mask (i)); -cpumask_set_cpu (i,cpu_core_mask (cpu)); -cpumask_set_cpu (cpu,cpu_core_mask (i)); -cpumask_set_cpu (i, c->llc_shared_map); -cpumask_set_cpu (cpu, o->llc_shared_map); +if (cpu_has (c,X86_FEATURE_TOPOEXT )) { +if (c->phys_proc_id == o->phys_proc_id && + c->compute_unit_id == o->compute_unit_id) +link_thread_siblings (cpu, i); + }else if (c->phys_proc_id == o->phys_proc_id && + c->cpu_core_id == o->cpu_core_id) { +link_thread_siblings (cpu, i); } } }else {
Signature · 1 line
1.6.4.4
[tip:x86/amd-nb] x86, amd: Use compute unit information to determine thread siblings
Commit-ID: d4fbe4f03557e1fd4d9bbb3a1957aad560f39e96 Gitweb: http://git.kernel.org/tip/d4fbe4f03557e1fd4d9bbb3a1957aad560f39e96 Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 30 Sep 2010 14:41:56 +0200 Committer: H. Peter Anvin <[email protected]> CommitDate: Fri, 1 Oct 2010 16:18:32 -0700
x86, amd: Use compute unit information to determine thread siblings
This information is vital for different load balancing policies.
LKML-Reference: <[email protected]>
--- arch/x86/kernel/smpboot.c | 28 ++++++++++++++++++++-------- 1 files changed, 20 insertions(+), 8 deletions(-)
arch/x86/kernel/smpboot.c +20 -8
@@ -397,6 +397,19 @@ void __cpuinit smp_store_cpu_info(int id)identify_secondary_cpu (c); } +static void __cpuinit link_thread_siblings (int cpu1,int cpu2) +{ +struct cpuinfo_x86 *c1 = &cpu_data (cpu1); +struct cpuinfo_x86 *c2 = &cpu_data (cpu2); + +cpumask_set_cpu (cpu1,cpu_sibling_mask (cpu2)); +cpumask_set_cpu (cpu2,cpu_sibling_mask (cpu1)); +cpumask_set_cpu (cpu1,cpu_core_mask (cpu2)); +cpumask_set_cpu (cpu2,cpu_core_mask (cpu1)); +cpumask_set_cpu (cpu1, c2->llc_shared_map); +cpumask_set_cpu (cpu2, c1->llc_shared_map); +} +void __cpuinit set_cpu_sibling_map (int cpu) {
@@ -409,14 +422,13 @@ void __cpuinit set_cpu_sibling_map(int cpu)for_each_cpu (i, cpu_sibling_setup_mask) {struct cpuinfo_x86 *o = &cpu_data (i); -if (c->phys_proc_id == o->phys_proc_id && - c->cpu_core_id == o->cpu_core_id) { -cpumask_set_cpu (i,cpu_sibling_mask (cpu)); -cpumask_set_cpu (cpu,cpu_sibling_mask (i)); -cpumask_set_cpu (i,cpu_core_mask (cpu)); -cpumask_set_cpu (cpu,cpu_core_mask (i)); -cpumask_set_cpu (i, c->llc_shared_map); -cpumask_set_cpu (cpu, o->llc_shared_map); +if (cpu_has (c,X86_FEATURE_TOPOEXT )) { +if (c->phys_proc_id == o->phys_proc_id && + c->compute_unit_id == o->compute_unit_id) +link_thread_siblings (cpu, i); + }else if (c->phys_proc_id == o->phys_proc_id && + c->cpu_core_id == o->cpu_core_id) { +link_thread_siblings (cpu, i); } } }else {
PATCH 6/6 x86, amd_nb: Enable GART support for AMD family 0x15 CPUs
In reply to Andreas HerrmannFrom: Andreas Herrmann <[email protected]>
AMD CPU family 0x15 still supports GART for compatibility reasons.
--- arch/x86/kernel/amd_nb.c | 4 +++- include/linux/pci_ids.h | 1 + 2 files changed, 4 insertions(+), 1 deletions(-)
arch/x86/kernel/amd_nb.c +3 -1
@@ -15,6 +15,7 @@ static u32 *flush_words;struct pci_device_id k8_nb_ids[] = { {PCI_DEVICE (PCI_VENDOR_ID_AMD ,PCI_DEVICE_ID_AMD_K8_NB_MISC ) }, {PCI_DEVICE (PCI_VENDOR_ID_AMD ,PCI_DEVICE_ID_AMD_10H_NB_MISC ) }, + {PCI_DEVICE (PCI_VENDOR_ID_AMD ,PCI_DEVICE_ID_AMD_15H_NB_MISC ) }, {} };EXPORT_SYMBOL (k8_nb_ids);
@@ -45,7 +46,8 @@ int cache_k8_northbridges(void) k8_northbridges.num++;/* some CPU families (e.g. family 0x11) do not support GART */ -if (boot_cpu_data.x86 ==0xf || boot_cpu_data.x86 ==0x10 ) +if (boot_cpu_data.x86 ==0xf || boot_cpu_data.x86 ==0x10 || + boot_cpu_data.x86 ==0x15 ) k8_northbridges.gart_supported =1 ; k8_northbridges.nb_misc = kmalloc((k8_northbridges.num + 1) *
include/linux/pci_ids.h +1 -0
@@ -517,6 +517,7 @@#define PCI_DEVICE_ID_AMD_11H_NB_DRAM 0x1302#define PCI_DEVICE_ID_AMD_11H_NB_MISC 0x1303#define PCI_DEVICE_ID_AMD_11H_NB_LINK 0x1304 +#define PCI_DEVICE_ID_AMD_15H_NB_MISC 0x1603#define PCI_DEVICE_ID_AMD_LANCE 0x2000#define PCI_DEVICE_ID_AMD_LANCE_HOME 0x2001#define PCI_DEVICE_ID_AMD_SCSI 0x2020
Signature · 1 line
1.6.4.4
[tip:x86/amd-nb] x86, amd_nb: Enable GART support for AMD family 0x15 CPUs
Commit-ID: 5c80cc78de46aef6cd5e714208da05c3f7f548f8 Gitweb: http://git.kernel.org/tip/5c80cc78de46aef6cd5e714208da05c3f7f548f8 Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 30 Sep 2010 14:43:16 +0200 Committer: H. Peter Anvin <[email protected]> CommitDate: Fri, 1 Oct 2010 16:18:32 -0700
x86, amd_nb: Enable GART support for AMD family 0x15 CPUs
AMD CPU family 0x15 still supports GART for compatibility reasons.
LKML-Reference: <[email protected]>
--- arch/x86/kernel/amd_nb.c | 4 +++- include/linux/pci_ids.h | 1 + 2 files changed, 4 insertions(+), 1 deletions(-)
arch/x86/kernel/amd_nb.c +3 -1
@@ -15,6 +15,7 @@ static u32 *flush_words;struct pci_device_id k8_nb_ids[] = { {PCI_DEVICE (PCI_VENDOR_ID_AMD ,PCI_DEVICE_ID_AMD_K8_NB_MISC ) }, {PCI_DEVICE (PCI_VENDOR_ID_AMD ,PCI_DEVICE_ID_AMD_10H_NB_MISC ) }, + {PCI_DEVICE (PCI_VENDOR_ID_AMD ,PCI_DEVICE_ID_AMD_15H_NB_MISC ) }, {} };EXPORT_SYMBOL (k8_nb_ids);
@@ -45,7 +46,8 @@ int cache_k8_northbridges(void) k8_northbridges.num++;/* some CPU families (e.g. family 0x11) do not support GART */ -if (boot_cpu_data.x86 ==0xf || boot_cpu_data.x86 ==0x10 ) +if (boot_cpu_data.x86 ==0xf || boot_cpu_data.x86 ==0x10 || + boot_cpu_data.x86 ==0x15 ) k8_northbridges.gart_supported =1 ; k8_northbridges.nb_misc = kmalloc((k8_northbridges.num + 1) *
include/linux/pci_ids.h +1 -0
@@ -514,6 +514,7 @@#define PCI_DEVICE_ID_AMD_11H_NB_DRAM 0x1302#define PCI_DEVICE_ID_AMD_11H_NB_MISC 0x1303#define PCI_DEVICE_ID_AMD_11H_NB_LINK 0x1304 +#define PCI_DEVICE_ID_AMD_15H_NB_MISC 0x1603#define PCI_DEVICE_ID_AMD_LANCE 0x2000#define PCI_DEVICE_ID_AMD_LANCE_HOME 0x2001#define PCI_DEVICE_ID_AMD_SCSI 0x2020