v5 multi-die/package topology support
Subsystems (9)
PATCH 9/22 powercap/intel_rapl: Simplify rapl_find_package()
In reply to Len BrownFrom: Zhang Rui <[email protected]>
Syntax only, no functional or semantic change.
Simplify how the code to discover a package is called. Rename find_package_by_id() to rapl_find_package_domain()
--- drivers/powercap/intel_rapl.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-)
drivers/powercap/intel_rapl.c +9 -9
@@ -264,8 +264,9 @@ static struct powercap_control_type *control_type; /* PowerCap Controller */static struct rapl_domain *platform_rapl_domain;/* Platform (PSys) domain */ /* caller to ensure CPU hotplug lock is held */ -static struct rapl_package *find_package_by_id (int id) +static struct rapl_package *rapl_find_package_domain (int cpu) { +int id =topology_physical_package_id (cpu);struct rapl_package *rp;list_for_each_entry (rp, &rapl_packages, plist) {
@@ -1300,7 +1301,7 @@ static int __init rapl_register_psys(void) rd->rpl[0 ].name = pl1_name; rd->rpl[1 ].prim_id =PL2_ENABLE ; rd->rpl[1 ].name = pl2_name; - rd->rp =find_package_by_id (0 ); + rd->rp =rapl_find_package_domain (0 ); power_zone = powercap_register_zone(&rd->power_zone, control_type, "psys", NULL,
@@ -1456,8 +1457,9 @@ static void rapl_remove_package(struct rapl_package *rp) }/* called from CPU hotplug notifier, hotplug lock held */ -static struct rapl_package *rapl_add_package (int cpu,int pkgid) +static struct rapl_package *rapl_add_package (int cpu) { +int id =topology_physical_package_id (cpu);struct rapl_package *rp;int ret;
@@ -1466,7 +1468,7 @@ static struct rapl_package *rapl_add_package(int cpu, int pkgid)return ERR_PTR (-ENOMEM );/* add the new package to the list */ - rp->id = pkgid; + rp->id = id; rp->lead_cpu = cpu;/* check if the package contains valid domains */
@@ -1497,12 +1499,11 @@ static struct rapl_package *rapl_add_package(int cpu, int pkgid)*/ static int rapl_cpu_online (unsigned int cpu) { -int pkgid =topology_physical_package_id (cpu);struct rapl_package *rp; - rp =find_package_by_id (pkgid); + rp =rapl_find_package_domain (cpu);if (!rp) { - rp =rapl_add_package (cpu, pkgid); + rp =rapl_add_package (cpu);if (IS_ERR (rp))return PTR_ERR (rp); }
@@ -1512,11 +1513,10 @@ static int rapl_cpu_online(unsigned int cpu)static int rapl_cpu_down_prep (unsigned int cpu) { -int pkgid =topology_physical_package_id (cpu);struct rapl_package *rp;int lead_cpu; - rp =find_package_by_id (pkgid); + rp =rapl_find_package_domain (cpu);if (!rp)return 0 ;
Signature · 1 line
2.18.0-rc0
PATCH 10/22 powercap/intel_rapl: Support multi-die/package
In reply to Len BrownFrom: Zhang Rui <[email protected]>
RAPL "package" domains are actually implemented in hardware per-die. Thus, the new multi-die/package systems have mulitple domains within each physical package.
Update the intel_rapl driver to be "die aware" -- exporting multiple domains within a single package, when present.
No change on single die/package systems.
--- drivers/powercap/intel_rapl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/powercap/intel_rapl.c +2 -2
@@ -266,7 +266,7 @@ static struct rapl_domain *platform_rapl_domain; /* Platform (PSys) domain *//* caller to ensure CPU hotplug lock is held */ static struct rapl_package *rapl_find_package_domain (int cpu) { -int id =topology_physical_package_id (cpu); +int id =topology_logical_die_id (cpu);struct rapl_package *rp;list_for_each_entry (rp, &rapl_packages, plist) {
@@ -1459,7 +1459,7 @@ static void rapl_remove_package(struct rapl_package *rp)/* called from CPU hotplug notifier, hotplug lock held */ static struct rapl_package *rapl_add_package (int cpu) { -int id =topology_physical_package_id (cpu); +int id =topology_logical_die_id (cpu);struct rapl_package *rp;int ret;
Signature · 1 line
2.18.0-rc0
PATCH 12/22 powercap/intel_rapl: update rapl domain name and debug messages
In reply to Len BrownFrom: Zhang Rui <[email protected]>
The RAPL domain "name" attribute contains "Package-N", which is ambiguous on multi-die per-package systems.
Update the name to "package-X-die-Y" on those systems.
No change on systems without multi-die/package.
Update driver debug messages.
--- drivers/powercap/intel_rapl.c | 57 ++++++++++++++++++++--------------- 1 file changed, 32 insertions(+), 25 deletions(-)
drivers/powercap/intel_rapl.c +32 -25
@@ -178,12 +178,15 @@ struct rapl_domain {#define power_zone_to_rapl_domain (_zone) \ container_of(_zone, struct rapl_domain, power_zone) +/* maximum rapl package domain name: package-%d-die-%d */ +#define PACKAGE_DOMAIN_NAME_LENGTH 30 -/* Each physical package contains multiple domains, these are the common + +/* Each rapl package contains multiple domains, these are the common * data across RAPL domains within a package. */ struct rapl_package { -unsigned int id;/* physical package/socket id */ +unsigned int id;/* logical die id, equals physical 1-die systems */ unsigned int nr_domains;unsigned long domain_map;/* bit map of active domains */ unsigned int power_unit;
@@ -198,6 +201,7 @@ struct rapl_package {int lead_cpu;/* one active cpu per package for access */ /* Track active cpus */ struct cpumask cpumask; +char name[PACKAGE_DOMAIN_NAME_LENGTH ]; };struct rapl_defaults {
@@ -926,8 +930,8 @@ static int rapl_check_unit_core(struct rapl_package *rp, int cpu) value = (msr_val &TIME_UNIT_MASK ) >>TIME_UNIT_OFFSET ; rp->time_unit =1000000 / (1 << value); -pr_debug ("Core CPU package %d energy=%dpJ, time=%dus, power=%duW\n" , - rp->id, rp->energy_unit, rp->time_unit, rp->power_unit); +pr_debug ("Core CPU %s energy=%dpJ, time=%dus, power=%duW\n" , + rp->name, rp->energy_unit, rp->time_unit, rp->power_unit);return 0 ; }
@@ -951,8 +955,8 @@ static int rapl_check_unit_atom(struct rapl_package *rp, int cpu) value = (msr_val &TIME_UNIT_MASK ) >>TIME_UNIT_OFFSET ; rp->time_unit =1000000 / (1 << value); -pr_debug ("Atom package %d energy=%dpJ, time=%dus, power=%duW\n" , - rp->id, rp->energy_unit, rp->time_unit, rp->power_unit); +pr_debug ("Atom %s energy=%dpJ, time=%dus, power=%duW\n" , + rp->name, rp->energy_unit, rp->time_unit, rp->power_unit);return 0 ; }
@@ -1181,7 +1185,7 @@ static void rapl_update_domain_data(struct rapl_package *rp)u64 val;for (dmn =0 ; dmn < rp->nr_domains; dmn++) { -pr_debug ("update package %d domain %s data\n" , rp->id, +pr_debug ("update %s domain %s data\n" , rp->name, rp->domains[dmn].name);/* exclude non-raw primitives */ for (prim =0 ; prim <NR_RAW_PRIMITIVES ; prim++) {
@@ -1206,7 +1210,6 @@ static void rapl_unregister_powercap(void)static int rapl_package_register_powercap (struct rapl_package *rp) {struct rapl_domain *rd; -char dev_name[17 ];/* max domain name = 7 + 1 + 8 for int + 1 for null*/ struct powercap_zone *power_zone =NULL ;int nr_pl, ret;
@@ -1217,20 +1220,16 @@ static int rapl_package_register_powercap(struct rapl_package *rp)for (rd = rp->domains; rd < rp->domains + rp->nr_domains; rd++) {if (rd->id ==RAPL_DOMAIN_PACKAGE ) { nr_pl =find_nr_power_limit (rd); -pr_debug ("register socket %d package domain %s\n" , - rp->id, rd->name); -memset (dev_name,0 ,sizeof (dev_name)); -snprintf (dev_name,sizeof (dev_name),"%s-%d" , - rd->name, rp->id); +pr_debug ("register package domain %s\n" , rp->name); power_zone =powercap_register_zone (&rd->power_zone, control_type, - dev_name,NULL , + rp->name,NULL , &zone_ops[rd->id], nr_pl, &constraint_ops);if (IS_ERR (power_zone)) { -pr_debug ("failed to register package, %d\n" , - rp->id); +pr_debug ("failed to register power zone %s\n" , + rp->name);return PTR_ERR (power_zone); }/* track parent zone in per package/socket data */
@@ -1256,8 +1255,8 @@ static int rapl_package_register_powercap(struct rapl_package *rp) &constraint_ops);if (IS_ERR (power_zone)) { -pr_debug ("failed to register power_zone, %d:%s:%s\n" , - rp->id, rd->name, dev_name); +pr_debug ("failed to register power_zone, %s:%s\n" , + rp->name, rd->name); ret =PTR_ERR (power_zone); goto err_cleanup; }
@@ -1270,7 +1269,7 @@ static int rapl_package_register_powercap(struct rapl_package *rp)* failed after the first domain setup. */ while (--rd >= rp->domains) { -pr_debug ("unregister package %d domain %s\n" , rp->id, rd->name); +pr_debug ("unregister %s domain %s\n" , rp->name, rd->name);powercap_unregister_zone (control_type, &rd->power_zone); }
@@ -1380,8 +1379,8 @@ static void rapl_detect_powerlimit(struct rapl_domain *rd)/* check if the domain is locked by BIOS, ignore if MSR doesn't exist */ if (!rapl_read_data_raw (rd,FW_LOCK , false, &val64)) {if (val64) { -pr_info ("RAPL package %d domain %s locked by BIOS\n" , - rd->rp->id, rd->name); +pr_info ("RAPL %s domain %s locked by BIOS\n" , + rd->rp->name, rd->name); rd->state |=DOMAIN_STATE_BIOS_LOCKED ; } }
@@ -1410,10 +1409,10 @@ static int rapl_detect_domains(struct rapl_package *rp, int cpu) } rp->nr_domains =bitmap_weight (&rp->domain_map,RAPL_DOMAIN_MAX );if (!rp->nr_domains) { -pr_debug ("no valid rapl domains found in package %d\n" , rp->id); +pr_debug ("no valid rapl domains found in %s\n" , rp->name);return -ENODEV ; } -pr_debug ("found %d domains on package %d\n" , rp->nr_domains, rp->id); +pr_debug ("found %d domains on %s\n" , rp->nr_domains, rp->name); rp->domains =kcalloc (rp->nr_domains +1 ,sizeof (struct rapl_domain ),GFP_KERNEL );
@@ -1446,8 +1445,8 @@ static void rapl_remove_package(struct rapl_package *rp) rd_package = rd;continue ; } -pr_debug ("remove package, undo power limit on %d: %s\n" , - rp->id, rd->name); +pr_debug ("remove package, undo power limit on %s: %s\n" , + rp->name, rd->name);powercap_unregister_zone (control_type, &rd->power_zone); }/* do parent zone last */
@@ -1461,6 +1460,7 @@ static struct rapl_package *rapl_add_package(int cpu) {int id =topology_logical_die_id (cpu);struct rapl_package *rp; +struct cpuinfo_x86 *c = &cpu_data (cpu);int ret; rp =kzalloc (sizeof (struct rapl_package ),GFP_KERNEL );
@@ -1471,6 +1471,13 @@ static struct rapl_package *rapl_add_package(int cpu) rp->id = id; rp->lead_cpu = cpu; +if (topology_max_die_per_package () >1 ) +snprintf (rp->name,PACKAGE_DOMAIN_NAME_LENGTH , +"package-%d-die-%d" , c->phys_proc_id, c->cpu_die_id); +else +snprintf (rp->name,PACKAGE_DOMAIN_NAME_LENGTH ,"package-%d" , + c->phys_proc_id); +/* check if the package contains valid domains */ if (rapl_detect_domains (rp, cpu) || rapl_defaults->check_unit (rp, cpu)) {
Signature · 1 line
2.18.0-rc0
PATCH 13/22 hwmon/coretemp: Support multi-die/package
In reply to Len BrownFrom: Zhang Rui <[email protected]>
Package temperature sensors are actually implemented in hardware per-die.
Update coretemp to be "die-aware", so it can expose mulitple sensors per package, instead of just one. No change to single-die/package systems.
--- drivers/hwmon/coretemp.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/hwmon/coretemp.c +4 -4
@@ -435,7 +435,7 @@ static int chk_ucode_version(unsigned int cpu)static struct platform_device *coretemp_get_pdev (unsigned int cpu) { -int pkgid =topology_logical_package_id (cpu); +int pkgid =topology_logical_die_id (cpu);if (pkgid >=0 && pkgid < max_packages)return pkg_devices[pkgid];
@@ -579,7 +579,7 @@ static struct platform_driver coretemp_driver = {static struct platform_device *coretemp_device_add (unsigned int cpu) { -int err, pkgid =topology_logical_package_id (cpu); +int err, pkgid =topology_logical_die_id (cpu);struct platform_device *pdev;if (pkgid <0 )
@@ -703,7 +703,7 @@ static int coretemp_cpu_offline(unsigned int cpu)* the rest. */ if (cpumask_empty (&pd->cpumask)) { - pkg_devices[topology_logical_package_id (cpu)] =NULL ; + pkg_devices[topology_logical_die_id (cpu)] =NULL ;platform_device_unregister (pdev);return 0 ; }
@@ -741,7 +741,7 @@ static int __init coretemp_init(void)if (!x86_match_cpu (coretemp_ids))return -ENODEV ; - max_packages =topology_max_packages (); + max_packages =topology_max_packages () *topology_max_die_per_package (); pkg_devices =kcalloc (max_packages,sizeof (struct platform_device *),GFP_KERNEL );if (!pkg_devices)
Signature · 1 line
2.18.0-rc0
cover v5 multi-die/package topology support
This patch series does 4 things.
1. Parse the new CPUID.1F leaf to discover multi-die/package topology
2. Export multi-die topology inside the kernel
3. Update 4 places (coretemp, pkgtemp, rapl, perf) that that need to know the difference between die and package-scope MSR.
4. Export multi-die topology to user-space via sysfs
These changes should have 0 impact on cache topology, NUMA topology, Linux scheduler, or system performance.
These topology changes primarily impact parts of the kernel and some applications that care about package MSR scope. Also, some software is licensed per package, and other tools, such as benchmark reporting software sometimes cares about packages.
---
Updates since v4
[PATCH 13/22] hwmon/coretemp: Support multi-die/package
Removed a dead line that should have gone away in the v3 cpuinfo_x86 cleanup suggested by tglx.
[PATCH 19/22] thermal/x86_pkg_temp_thermal: rename internal variables to zones from packages [PATCH 20/22] hwmon/coretemp: rename internal variables to zones from packages [PATCH 21/22] perf/x86/intel/uncore: renames in response to multi-die/pkg support [PATCH 22/22] perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support
New syntax-only patches to clean up in-consistent variable names resulting from previous patches. Suggested by ingo.
---
Updates since v2:
In response to brice, peterz and Morten Rasmussen,
used the word "cpu" rather than "thread" for the new sysfs attributes.
In response to tglx, replaced access to cpuinfo_x86.x86_max_dies, with macro topology_max_die_per_package(). In doing so, deleted this new per-cpu field entirely, as a global is sufficient.
Also, appended 3 patches from Kan Liang, updating the perf code to be multi-die aware. These patches are similar to the preceding power and temperature patches. I believe that with these patches, this series now includes all needed multi-die kernel support.
---
The following changes since commit 085b7755808aa11f78ab9377257e1dad2e6fa4bb:
Linux 5.1-rc6 (2019-04-21 10:45:57 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux.git x86
for you to fetch changes up to 9f57786ba08d4d5e913cd21693aadb0ccdba72b2:
perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support (2019-05-06 17:17:58 -0400)
----------------------------------------------------------------
Kan Liang (3):
perf/x86/intel/uncore: Support multi-die/package perf/x86/intel/rapl: Support multi-die/package perf/x86/intel/cstate: Support multi-die/package
Len Brown (14):
x86 topology: Fix doc typo
topology: Simplify cputopology.txt formatting and wording
x86 smpboot: Rename match_die() to match_pkg()
x86 topology: Add CPUID.1F multi-die/package support
x86 topology: Create topology_max_die_per_package()
cpu topology: Export die_id
x86 topology: Define topology_die_id()
x86 topology: Define topology_logical_die_id()
topology: Create package_cpus sysfs attribute
topology: Create core_cpus and die_cpus sysfs attributes thermal/x86_pkg_temp_thermal: rename internal variables to zones from packages hwmon/coretemp: rename internal variables to zones from packages perf/x86/intel/uncore: renames in response to multi-die/pkg support perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support
Zhang Rui (5):
powercap/intel_rapl: Simplify rapl_find_package() powercap/intel_rapl: Support multi-die/package thermal/x86_pkg_temp_thermal: Support multi-die/package powercap/intel_rapl: update rapl domain name and debug messages hwmon/coretemp: Support multi-die/package
Documentation/cputopology.txt | 80 +++++++++------ Documentation/x86/topology.txt | 6 +- arch/x86/events/intel/cstate.c | 14 ++- arch/x86/events/intel/rapl.c | 12 +-- arch/x86/events/intel/uncore.c | 80 +++++++-------- arch/x86/events/intel/uncore.h | 4 +- arch/x86/events/intel/uncore_snbep.c | 2 +- arch/x86/include/asm/processor.h | 4 +- arch/x86/include/asm/smp.h | 1 + arch/x86/include/asm/topology.h | 17 ++++ arch/x86/kernel/cpu/common.c | 1 + arch/x86/kernel/cpu/topology.c | 88 +++++++++++++---- arch/x86/kernel/smpboot.c | 75 +++++++++++++- arch/x86/xen/smp_pv.c | 1 + drivers/base/topology.c | 22 +++++ drivers/hwmon/coretemp.c | 36 +++---- drivers/powercap/intel_rapl.c | 75 +++++++------- drivers/thermal/intel/x86_pkg_temp_thermal.c | 142 ++++++++++++++------------- include/linux/topology.h | 6 ++ 19 files changed, 437 insertions(+), 229 deletions(-)
PATCH 1/22 x86 topology: Fix doc typo
From: Len Brown <[email protected]>
Syntax only, no functional or semantic change.
reflect actual cpuinfo_x86 field name:
s/logical_id/logical_proc_id/
--- Documentation/x86/topology.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Documentation/x86/topology.txt +1 -1
@@ -51,7 +51,7 @@ The topology of a system is described in the units of: The physical ID of the package. This information is retrieved via CPUID and deduced from the APIC IDs of the cores in the package. - - cpuinfo_x86.logical_id: + - cpuinfo_x86.logical_proc_id: The logical ID of the package. As we do not trust BIOSes to enumerate the packages in a consistent way, we introduced the concept of logical package
Signature · 1 line
2.18.0-rc0
PATCH 4/22 x86 topology: Add CPUID.1F multi-die/package support
From: Len Brown <[email protected]>
Some new systems have multiple software-visible die within each package.
Update Linux parsing of the Intel CPUID "Extended Topology Leaf" to handle either CPUID.B, or the new CPUID.1F.
Add cpuinfo_x86.die_id and cpuinfo_x86.max_dies to store the result.
die_id will be non-zero only for multi-die/package systems.
--- Documentation/x86/topology.txt | 4 ++ arch/x86/include/asm/processor.h | 4 +- arch/x86/kernel/cpu/topology.c | 85 +++++++++++++++++++++++++------- arch/x86/kernel/smpboot.c | 2 + 4 files changed, 75 insertions(+), 20 deletions(-)
Documentation/x86/topology.txt +4 -0
@@ -46,6 +46,10 @@ The topology of a system is described in the units of: The number of cores in a package. This information is retrieved via CPUID. + - cpuinfo_x86.x86_max_dies: + + The number of dies in a package. This information is retrieved via CPUID. + - cpuinfo_x86.phys_proc_id: The physical ID of the package. This information is retrieved via CPUID
arch/x86/include/asm/processor.h +3 -1
@@ -105,7 +105,8 @@ struct cpuinfo_x86 {int x86_power;unsigned long loops_per_jiffy;/* cpuid returned max cores value: */ -u16 x86_max_cores; +u16 x86_max_cores; +u16 x86_max_dies;u16 apicid;u16 initial_apicid;u16 x86_clflush_size;
@@ -117,6 +118,7 @@ struct cpuinfo_x86 {u16 logical_proc_id;/* Core id: */ u16 cpu_core_id; +u16 cpu_die_id;/* Index into per_cpu list: */ u16 cpu_index;u32 microcode;
arch/x86/kernel/cpu/topology.c +66 -19
151 lines not on this page · raw message
arch/x86/kernel/smpboot.c +2 -0
@@ -389,6 +389,7 @@ static bool match_smt(struct cpuinfo_x86 *c, struct cpuinfo_x86 *o)int cpu1 = c->cpu_index, cpu2 = o->cpu_index;if (c->phys_proc_id == o->phys_proc_id && + c->cpu_die_id == o->cpu_die_id &&per_cpu (cpu_llc_id, cpu1) ==per_cpu (cpu_llc_id, cpu2)) {if (c->cpu_core_id == o->cpu_core_id)return topology_sane (c, o,"smt" );
@@ -400,6 +401,7 @@ static bool match_smt(struct cpuinfo_x86 *c, struct cpuinfo_x86 *o) } }else if (c->phys_proc_id == o->phys_proc_id && + c->cpu_die_id == o->cpu_die_id && c->cpu_core_id == o->cpu_core_id) {return topology_sane (c, o,"smt" ); }
Signature · 1 line
2.18.0-rc0
PATCH 6/22 cpu topology: Export die_id
In reply to Len BrownFrom: Len Brown <[email protected]>
Export die_id in cpu topology, for the benefit of hardware that has multiple-die/package.
--- Documentation/cputopology.txt | 15 ++++++++++++--- drivers/base/topology.c | 4 ++++ include/linux/topology.h | 3 +++ 3 files changed, 19 insertions(+), 3 deletions(-)
Documentation/cputopology.txt +12 -3
@@ -12,6 +12,12 @@ physical_package_id: socket number, but the actual value is architecture and platform dependent. +die_id: + + the CPU die ID of cpuX. Typically it is the hardware platform's + identifier (rather than the kernel's). The actual value is + architecture and platform dependent. + core_id: the CPU core ID of cpuX. Typically it is the hardware platform's
@@ -81,6 +87,7 @@ For an architecture to support this feature, it must define some of these macros in include/asm-XXX/topology.h:: #define topology_physical_package_id(cpu) + #define topology_die_id(cpu) #define topology_core_id(cpu) #define topology_book_id(cpu) #define topology_drawer_id(cpu)
@@ -99,9 +106,11 @@ provides default definitions for any of the above macros that are not defined by include/asm-XXX/topology.h: 1) topology_physical_package_id: -1 -2) topology_core_id: 0 -3) topology_sibling_cpumask: just the given CPU -4) topology_core_cpumask: just the given CPU +2) topology_die_id: -1 +3) topology_core_id: 0 +4) topology_sibling_cpumask: just the given CPU +5) topology_core_cpumask: just the given CPU +6) topology_die_cpumask: just the given CPU For architectures that don't support books (CONFIG_SCHED_BOOK) there are no default definitions for topology_book_id() and topology_book_cpumask().
drivers/base/topology.c +4 -0
@@ -43,6 +43,9 @@ static ssize_t name##_list_show(struct device *dev, \define_id_show_func (physical_package_id);static DEVICE_ATTR_RO (physical_package_id); +define_id_show_func (die_id); +static DEVICE_ATTR_RO (die_id); +define_id_show_func (core_id);static DEVICE_ATTR_RO (core_id);
@@ -72,6 +75,7 @@ static DEVICE_ATTR_RO(drawer_siblings_list);static struct attribute *default_attrs[] = { &dev_attr_physical_package_id.attr, + &dev_attr_die_id.attr, &dev_attr_core_id.attr, &dev_attr_thread_siblings.attr, &dev_attr_thread_siblings_list.attr,
include/linux/topology.h +3 -0
@@ -184,6 +184,9 @@ static inline int cpu_to_mem(int cpu)#ifndef topology_physical_package_id#define topology_physical_package_id (cpu) ((void)(cpu), -1)#endif +#ifndef topology_die_id +#define topology_die_id (cpu) ((void)(cpu), -1) +#endif #ifndef topology_core_id#define topology_core_id (cpu) ((void)(cpu), 0)#endif
Signature · 1 line
2.18.0-rc0
PATCH 8/22 x86 topology: Define topology_logical_die_id()
In reply to Len BrownFrom: Len Brown <[email protected]>
Define topology_logical_die_id() ala
existing topology_logical_package_id()
--- arch/x86/include/asm/processor.h | 1 + arch/x86/include/asm/topology.h | 5 ++++ arch/x86/kernel/cpu/common.c | 1 + arch/x86/kernel/smpboot.c | 45 ++++++++++++++++++++++++++++++++ 4 files changed, 52 insertions(+)
arch/x86/include/asm/processor.h +1 -0
@@ -118,6 +118,7 @@ struct cpuinfo_x86 {/* Core id: */ u16 cpu_core_id;u16 cpu_die_id; +u16 logical_die_id;/* Index into per_cpu list: */ u16 cpu_index;u32 microcode;
arch/x86/include/asm/topology.h +5 -0
@@ -106,6 +106,7 @@ extern const struct cpumask *cpu_coregroup_mask(int cpu);#define topology_logical_package_id (cpu) (cpu_data(cpu).logical_proc_id)#define topology_physical_package_id (cpu) (cpu_data(cpu).phys_proc_id) +#define topology_logical_die_id (cpu) (cpu_data(cpu).logical_die_id)#define topology_die_id (cpu) (cpu_data(cpu).cpu_die_id)#define topology_core_id (cpu) (cpu_data(cpu).cpu_core_id)
@@ -131,13 +132,17 @@ static inline int topology_max_smt_threads(void) }int topology_update_package_map (unsigned int apicid,unsigned int cpu); +int topology_update_die_map (unsigned int dieid,unsigned int cpu);int topology_phys_to_logical_pkg (unsigned int pkg); +int topology_phys_to_logical_die (unsigned int die,unsigned int cpu);bool topology_is_primary_thread (unsigned int cpu);bool topology_smt_supported (void );#else #define topology_max_packages () (1)static inline int topology_update_package_map (unsigned int apicid,unsigned int cpu) {return 0 ; } +static inline int +topology_update_die_map (unsigned int dieid,unsigned int cpu) {return 0 ; }static inline int topology_phys_to_logical_pkg (unsigned int pkg) {return 0 ; }static inline int topology_phys_to_logical_die (unsigned int die,unsigned int cpu) {return 0 ; }
arch/x86/kernel/cpu/common.c +1 -0
@@ -1285,6 +1285,7 @@ static void validate_apic_and_package_id(struct cpuinfo_x86 *c) cpu, apicid, c->initial_apicid); }BUG_ON (topology_update_package_map (c->phys_proc_id, cpu)); +BUG_ON (topology_update_die_map (c->cpu_die_id, cpu));#else c->logical_proc_id =0 ;#endif
arch/x86/kernel/smpboot.c +45 -0
69 lines not on this page · raw message
Signature · 1 line
2.18.0-rc0
PATCH 16/22 perf/x86/intel/uncore: Support multi-die/package
In reply to Len BrownFrom: Kan Liang <[email protected]>
Uncore becomes die-scope on Xeon Cascade Lake-AP. Uncore driver needs to support die-scope uncore units.
Use topology_logical_die_id() to replace topology_logical_package_id(). For previous platforms which doesn't have multi-die, topology_logical_die_id() is identical as topology_logical_package_id().
In pci_probe/remove, the group id reads from PCI BUS is logical die id for multi-die systems.
Use topology_die_cpumask() to replace topology_core_cpumask(). For previous platforms which doesn't have multi-die, topology_die_cpumask() is identical as topology_core_cpumask().
There is no functional change for previous platforms.
--- arch/x86/events/intel/uncore.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-)
arch/x86/events/intel/uncore.c +11 -9
@@ -100,7 +100,7 @@ ssize_t uncore_event_show(struct kobject *kobj,struct intel_uncore_box *uncore_pmu_to_box (struct intel_uncore_pmu *pmu,int cpu) { -unsigned int pkgid =topology_logical_package_id (cpu); +unsigned int pkgid =topology_logical_die_id (cpu);/* * The unsigned check also catches the '-1' return value for non
@@ -942,7 +942,8 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *idif (phys_id <0 )return -ENODEV ; - pkg =topology_phys_to_logical_pkg (phys_id); + pkg = (topology_max_die_per_package () >1 ) ? phys_id : +topology_phys_to_logical_pkg (phys_id);if (pkg <0 )return -EINVAL ;
@@ -1033,7 +1034,8 @@ static void uncore_pci_remove(struct pci_dev *pdev) box =pci_get_drvdata (pdev);if (!box) { - pkg =topology_phys_to_logical_pkg (phys_id); + pkg = (topology_max_die_per_package () >1 ) ? phys_id : +topology_phys_to_logical_pkg (phys_id);for (i =0 ; i <UNCORE_EXTRA_PCI_DEV_MAX ; i++) {if (uncore_extra_pci_dev[pkg].dev[i] == pdev) { uncore_extra_pci_dev[pkg].dev[i] =NULL ;
@@ -1110,7 +1112,7 @@ static void uncore_change_type_ctx(struct intel_uncore_type *type, int old_cpu,struct intel_uncore_box *box;int i, pkg; - pkg =topology_logical_package_id (old_cpu <0 ? new_cpu : old_cpu); + pkg =topology_logical_die_id (old_cpu <0 ? new_cpu : old_cpu);for (i =0 ; i < type->num_boxes; i++, pmu++) { box = pmu->boxes[pkg];if (!box)
@@ -1151,7 +1153,7 @@ static int uncore_event_cpu_offline(unsigned int cpu)if (!cpumask_test_and_clear_cpu (cpu, &uncore_cpu_mask)) goto unref;/* Find a new cpu to collect uncore events */ - target =cpumask_any_but (topology_core_cpumask (cpu), cpu); + target =cpumask_any_but (topology_die_cpumask (cpu), cpu);/* Migrate uncore events to the new target */ if (target < nr_cpu_ids)
@@ -1164,7 +1166,7 @@ static int uncore_event_cpu_offline(unsigned int cpu) unref:/* Clear the references */ - pkg =topology_logical_package_id (cpu); + pkg =topology_logical_die_id (cpu);for (; *types; types++) { type = *types; pmu = type->pmus;
@@ -1223,7 +1225,7 @@ static int uncore_event_cpu_online(unsigned int cpu)struct intel_uncore_box *box;int i, ret, pkg, target; - pkg =topology_logical_package_id (cpu); + pkg =topology_logical_die_id (cpu); ret =allocate_boxes (types, pkg, cpu);if (ret)return ret;
@@ -1242,7 +1244,7 @@ static int uncore_event_cpu_online(unsigned int cpu)* Check if there is an online cpu in the package * which collects uncore events already. */ - target =cpumask_any_and (&uncore_cpu_mask,topology_core_cpumask (cpu)); + target =cpumask_any_and (&uncore_cpu_mask,topology_die_cpumask (cpu));if (target < nr_cpu_ids)return 0 ;
@@ -1411,7 +1413,7 @@ static int __init intel_uncore_init(void)if (boot_cpu_has (X86_FEATURE_HYPERVISOR ))return -ENODEV ; - max_packages =topology_max_packages (); + max_packages =topology_max_packages () *topology_max_die_per_package (); uncore_init = (struct intel_uncore_init_fun *)id->driver_data;if (uncore_init->pci_init) {
Signature · 1 line
2.18.0-rc0
PATCH 18/22 perf/x86/intel/cstate: Support multi-die/package
In reply to Len BrownFrom: Kan Liang <[email protected]>
Some cstate counters becomes die-scope on Xeon Cascade Lake-AP. Perf cstate driver needs to support die-scope cstate counters.
Use topology_die_cpumask() to replace topology_core_cpumask(). For previous platforms which doesn't have multi-die, topology_die_cpumask() is identical as topology_core_cpumask(). There is no functional change for previous platforms.
Name the die-scope PMU "cstate_die".
--- arch/x86/events/intel/cstate.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-)
arch/x86/events/intel/cstate.c +10 -4
@@ -302,7 +302,7 @@ static int cstate_pmu_event_init(struct perf_event *event)return -EINVAL ; event->hw.event_base = pkg_msr[cfg].msr; cpu =cpumask_any_and (&cstate_pkg_cpu_mask, -topology_core_cpumask (event->cpu)); +topology_die_cpumask (event->cpu)); }else {return -ENOENT ; }
@@ -385,7 +385,7 @@ static int cstate_cpu_exit(unsigned int cpu)if (has_cstate_pkg &&cpumask_test_and_clear_cpu (cpu, &cstate_pkg_cpu_mask)) { - target =cpumask_any_but (topology_core_cpumask (cpu), cpu); + target =cpumask_any_but (topology_die_cpumask (cpu), cpu);/* Migrate events if there is a valid target */ if (target < nr_cpu_ids) {cpumask_set_cpu (target, &cstate_pkg_cpu_mask);
@@ -414,7 +414,7 @@ static int cstate_cpu_init(unsigned int cpu)* in the package cpu mask as the designated reader. */ target =cpumask_any_and (&cstate_pkg_cpu_mask, -topology_core_cpumask (cpu)); +topology_die_cpumask (cpu));if (has_cstate_pkg && target >= nr_cpu_ids)cpumask_set_cpu (cpu, &cstate_pkg_cpu_mask);
@@ -661,7 +661,13 @@ static int __init cstate_init(void) }if (has_cstate_pkg) { - err =perf_pmu_register (&cstate_pkg_pmu, cstate_pkg_pmu.name,-1 ); +if (topology_max_die_per_package () >1 ) { + err =perf_pmu_register (&cstate_pkg_pmu, +"cstate_die" ,-1 ); + }else { + err =perf_pmu_register (&cstate_pkg_pmu, + cstate_pkg_pmu.name,-1 ); + }if (err) { has_cstate_pkg = false;pr_info ("Failed to register cstate pkg pmu\n" );
Signature · 1 line
2.18.0-rc0
PATCH 17/22 perf/x86/intel/rapl: Support multi-die/package
In reply to Len BrownFrom: Kan Liang <[email protected]>
RAPL becomes die-scope on Xeon Cascade Lake-AP. Perf RAPL driver needs to support die-scope RAPL domain.
Use topology_logical_die_id() to replace topology_logical_package_id(). For previous platforms which doesn't have multi-die, topology_logical_die_id() is identical as topology_logical_package_id().
Use topology_die_cpumask() to replace topology_core_cpumask(). For previous platforms which doesn't have multi-die, topology_die_cpumask() is identical as topology_core_cpumask().
There is no functional change for previous platforms.
--- arch/x86/events/intel/rapl.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
arch/x86/events/intel/rapl.c +5 -5
@@ -161,7 +161,7 @@ static u64 rapl_timer_ms;static inline struct rapl_pmu *cpu_to_rapl_pmu (unsigned int cpu) { -unsigned int pkgid =topology_logical_package_id (cpu); +unsigned int pkgid =topology_logical_die_id (cpu);/* * The unsigned check also catches the '-1' return value for non
@@ -571,7 +571,7 @@ static int rapl_cpu_offline(unsigned int cpu) pmu->cpu =-1 ;/* Find a new cpu to collect rapl events */ - target =cpumask_any_but (topology_core_cpumask (cpu), cpu); + target =cpumask_any_but (topology_die_cpumask (cpu), cpu);/* Migrate rapl events to the new target */ if (target < nr_cpu_ids) {
@@ -598,14 +598,14 @@ static int rapl_cpu_online(unsigned int cpu) pmu->timer_interval =ms_to_ktime (rapl_timer_ms);rapl_hrtimer_init (pmu); - rapl_pmus->pmus[topology_logical_package_id (cpu)] = pmu; + rapl_pmus->pmus[topology_logical_die_id (cpu)] = pmu; }/* * Check if there is an online cpu in the package which collects rapl * events already. */ - target =cpumask_any_and (&rapl_cpu_mask,topology_core_cpumask (cpu)); + target =cpumask_any_and (&rapl_cpu_mask,topology_die_cpumask (cpu));if (target < nr_cpu_ids)return 0 ;
@@ -675,7 +675,7 @@ static void cleanup_rapl_pmus(void)static int __init init_rapl_pmus (void ) { -int maxpkg =topology_max_packages (); +int maxpkg =topology_max_packages () *topology_max_die_per_package ();size_t size; size =sizeof (*rapl_pmus) + maxpkg *sizeof (struct rapl_pmu *);
Signature · 1 line
2.18.0-rc0
PATCH 21/22 perf/x86/intel/uncore: renames in response to multi-die/pkg support
In reply to Len BrownFrom: Len Brown <[email protected]>
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use the more generic "box" terminology, instead of "pkg", as the boxes can refer to either packages or die.
--- arch/x86/events/intel/uncore.c | 74 ++++++++++++++-------------- arch/x86/events/intel/uncore.h | 4 +- arch/x86/events/intel/uncore_snbep.c | 2 +- 3 files changed, 40 insertions(+), 40 deletions(-)
arch/x86/events/intel/uncore.c +37 -37
@@ -14,7 +14,7 @@ struct pci_driver *uncore_pci_driver;DEFINE_RAW_SPINLOCK (pci2phy_map_lock);struct list_head pci2phy_map_head =LIST_HEAD_INIT (pci2phy_map_head);struct pci_extra_dev *uncore_extra_pci_dev; -static int max_packages; +static int max_boxes;/* mask of cpus that collect uncore events */ static cpumask_t uncore_cpu_mask;
@@ -100,13 +100,13 @@ ssize_t uncore_event_show(struct kobject *kobj,struct intel_uncore_box *uncore_pmu_to_box (struct intel_uncore_pmu *pmu,int cpu) { -unsigned int pkgid =topology_logical_die_id (cpu); +unsigned int boxid =topology_logical_die_id (cpu);/* * The unsigned check also catches the '-1' return value for non * existent mappings in the topology map. */ -return pkgid < max_packages ? pmu->boxes[pkgid] :NULL ; +return boxid < max_boxes ? pmu->boxes[boxid] :NULL ; }u64 uncore_msr_read_counter (struct intel_uncore_box *box,struct perf_event *event)
@@ -311,7 +311,7 @@ static struct intel_uncore_box *uncore_alloc_box(struct intel_uncore_type *type,uncore_pmu_init_hrtimer (box); box->cpu =-1 ; box->pci_phys_id =-1 ; - box->pkgid =-1 ; + box->boxid =-1 ;/* set default hrtimer timeout */ box->hrtimer_duration =UNCORE_PMU_HRTIMER_INTERVAL ;
@@ -826,10 +826,10 @@ static void uncore_pmu_unregister(struct intel_uncore_pmu *pmu)static void uncore_free_boxes (struct intel_uncore_pmu *pmu) { -int pkg; +int boxid; -for (pkg =0 ; pkg < max_packages; pkg++) -kfree (pmu->boxes[pkg]); +for (boxid =0 ; boxid < max_boxes; boxid++) +kfree (pmu->boxes[boxid]);kfree (pmu->boxes); }
@@ -866,7 +866,7 @@ static int __init uncore_type_init(struct intel_uncore_type *type, bool setid)if (!pmus)return -ENOMEM ; - size = max_packages *sizeof (struct intel_uncore_box *); + size = max_boxes *sizeof (struct intel_uncore_box *);for (i =0 ; i < type->num_boxes; i++) { pmus[i].func_id = setid ? i :-1 ;
@@ -936,21 +936,21 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *idstruct intel_uncore_type *type;struct intel_uncore_pmu *pmu =NULL ;struct intel_uncore_box *box; -int phys_id, pkg, ret; +int phys_id, boxid, ret; phys_id =uncore_pcibus_to_physid (pdev->bus);if (phys_id <0 )return -ENODEV ; - pkg = (topology_max_die_per_package () >1 ) ? phys_id : + boxid = (topology_max_die_per_package () >1 ) ? phys_id :topology_phys_to_logical_pkg (phys_id); -if (pkg <0 ) +if (boxid <0 )return -EINVAL ;if (UNCORE_PCI_DEV_TYPE (id->driver_data) ==UNCORE_EXTRA_PCI_DEV ) {int idx =UNCORE_PCI_DEV_IDX (id->driver_data); - uncore_extra_pci_dev[pkg].dev[idx] = pdev; + uncore_extra_pci_dev[boxid].dev[idx] = pdev;pci_set_drvdata (pdev,NULL );return 0 ; }
@@ -989,7 +989,7 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id pmu = &type->pmus[UNCORE_PCI_DEV_IDX (id->driver_data)]; } -if (WARN_ON_ONCE (pmu->boxes[pkg] !=NULL )) +if (WARN_ON_ONCE (pmu->boxes[boxid] !=NULL ))return -EINVAL ; box =uncore_alloc_box (type,NUMA_NO_NODE );
@@ -1003,13 +1003,13 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *idatomic_inc (&box->refcnt); box->pci_phys_id = phys_id; - box->pkgid = pkg; + box->boxid = boxid; box->pci_dev = pdev; box->pmu = pmu;uncore_box_init (box);pci_set_drvdata (pdev, box); - pmu->boxes[pkg] = box; + pmu->boxes[boxid] = box;if (atomic_inc_return (&pmu->activeboxes) >1 )return 0 ;
@@ -1017,7 +1017,7 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id ret =uncore_pmu_register (pmu);if (ret) {pci_set_drvdata (pdev,NULL ); - pmu->boxes[pkg] =NULL ; + pmu->boxes[boxid] =NULL ;uncore_box_exit (box);kfree (box); }
@@ -1028,17 +1028,17 @@ static void uncore_pci_remove(struct pci_dev *pdev) {struct intel_uncore_box *box;struct intel_uncore_pmu *pmu; -int i, phys_id, pkg; +int i, phys_id, boxid; phys_id =uncore_pcibus_to_physid (pdev->bus); box =pci_get_drvdata (pdev);if (!box) { - pkg = (topology_max_die_per_package () >1 ) ? phys_id : + boxid = (topology_max_die_per_package () >1 ) ? phys_id :topology_phys_to_logical_pkg (phys_id);for (i =0 ; i <UNCORE_EXTRA_PCI_DEV_MAX ; i++) { -if (uncore_extra_pci_dev[pkg].dev[i] == pdev) { - uncore_extra_pci_dev[pkg].dev[i] =NULL ; +if (uncore_extra_pci_dev[boxid].dev[i] == pdev) { + uncore_extra_pci_dev[boxid].dev[i] =NULL ;break ; } }
@@ -1051,7 +1051,7 @@ static void uncore_pci_remove(struct pci_dev *pdev)return ;pci_set_drvdata (pdev,NULL ); - pmu->boxes[box->pkgid] =NULL ; + pmu->boxes[box->boxid] =NULL ;if (atomic_dec_return (&pmu->activeboxes) ==0 )uncore_pmu_unregister (pmu);uncore_box_exit (box);
@@ -1063,7 +1063,7 @@ static int __init uncore_pci_init(void)size_t size;int ret; - size = max_packages *sizeof (struct pci_extra_dev ); + size = max_boxes *sizeof (struct pci_extra_dev ); uncore_extra_pci_dev =kzalloc (size,GFP_KERNEL );if (!uncore_extra_pci_dev) { ret = -ENOMEM ;
@@ -1110,11 +1110,11 @@ static void uncore_change_type_ctx(struct intel_uncore_type *type, int old_cpu, {struct intel_uncore_pmu *pmu = type->pmus;struct intel_uncore_box *box; -int i, pkg; +int i, boxid; - pkg =topology_logical_die_id (old_cpu <0 ? new_cpu : old_cpu); + boxid =topology_logical_die_id (old_cpu <0 ? new_cpu : old_cpu);for (i =0 ; i < type->num_boxes; i++, pmu++) { - box = pmu->boxes[pkg]; + box = pmu->boxes[boxid];if (!box)continue ;
@@ -1147,7 +1147,7 @@ static int uncore_event_cpu_offline(unsigned int cpu)struct intel_uncore_type *type, **types = uncore_msr_uncores;struct intel_uncore_pmu *pmu;struct intel_uncore_box *box; -int i, pkg, target; +int i, boxid, target;/* Check if exiting cpu is used for collecting uncore events */ if (!cpumask_test_and_clear_cpu (cpu, &uncore_cpu_mask))
@@ -1166,12 +1166,12 @@ static int uncore_event_cpu_offline(unsigned int cpu) unref:/* Clear the references */ - pkg =topology_logical_die_id (cpu); + boxid =topology_logical_die_id (cpu);for (; *types; types++) { type = *types; pmu = type->pmus;for (i =0 ; i < type->num_boxes; i++, pmu++) { - box = pmu->boxes[pkg]; + box = pmu->boxes[boxid];if (box &&atomic_dec_return (&box->refcnt) ==0 )uncore_box_exit (box); }
@@ -1180,7 +1180,7 @@ static int uncore_event_cpu_offline(unsigned int cpu) }static int allocate_boxes (struct intel_uncore_type **types, -unsigned int pkg,unsigned int cpu) +unsigned int boxid,unsigned int cpu) {struct intel_uncore_box *box, *tmp;struct intel_uncore_type *type;
@@ -1193,20 +1193,20 @@ static int allocate_boxes(struct intel_uncore_type **types, type = *types; pmu = type->pmus;for (i =0 ; i < type->num_boxes; i++, pmu++) { -if (pmu->boxes[pkg]) +if (pmu->boxes[boxid])continue ; box =uncore_alloc_box (type,cpu_to_node (cpu));if (!box) goto cleanup; box->pmu = pmu; - box->pkgid = pkg; + box->boxid = boxid;list_add (&box->active_list, &allocated); } }/* Install them in the pmus */ list_for_each_entry_safe (box, tmp, &allocated, active_list) {list_del_init (&box->active_list); - box->pmu->boxes[pkg] = box; + box->pmu->boxes[boxid] = box; }return 0 ;
@@ -1223,10 +1223,10 @@ static int uncore_event_cpu_online(unsigned int cpu)struct intel_uncore_type *type, **types = uncore_msr_uncores;struct intel_uncore_pmu *pmu;struct intel_uncore_box *box; -int i, ret, pkg, target; +int i, ret, boxid, target; - pkg =topology_logical_die_id (cpu); - ret =allocate_boxes (types, pkg, cpu); + boxid =topology_logical_die_id (cpu); + ret =allocate_boxes (types, boxid, cpu);if (ret)return ret;
@@ -1234,7 +1234,7 @@ static int uncore_event_cpu_online(unsigned int cpu) type = *types; pmu = type->pmus;for (i =0 ; i < type->num_boxes; i++, pmu++) { - box = pmu->boxes[pkg]; + box = pmu->boxes[boxid];if (box &&atomic_inc_return (&box->refcnt) ==1 )uncore_box_init (box); }
@@ -1413,7 +1413,7 @@ static int __init intel_uncore_init(void)if (boot_cpu_has (X86_FEATURE_HYPERVISOR ))return -ENODEV ; - max_packages =topology_max_packages () *topology_max_die_per_package (); + max_boxes =topology_max_packages () *topology_max_die_per_package (); uncore_init = (struct intel_uncore_init_fun *)id->driver_data;if (uncore_init->pci_init) {
arch/x86/events/intel/uncore.h +2 -2
@@ -108,7 +108,7 @@ struct intel_uncore_extra_reg {struct intel_uncore_box {int pci_phys_id; -int pkgid;/* Logical package ID */ +int boxid;/* Logical box ID */ int n_active;/* number of active events */ int n_events;int cpu;/* cpu to collect events */
@@ -467,7 +467,7 @@ static inline void uncore_box_exit(struct intel_uncore_box *box)static inline bool uncore_box_is_fake (struct intel_uncore_box *box) { -return (box->pkgid <0 ); +return (box->boxid <0 ); }static inline struct intel_uncore_pmu *uncore_event_to_pmu (struct perf_event *event)
arch/x86/events/intel/uncore_snbep.c +1 -1
@@ -1058,7 +1058,7 @@ static void snbep_qpi_enable_event(struct intel_uncore_box *box, struct perf_eveif (reg1->idx !=EXTRA_REG_NONE ) {int idx = box->pmu->pmu_idx +SNBEP_PCI_QPI_PORT0_FILTER ; -int pkg = box->pkgid; +int pkg = box->boxid;struct pci_dev *filter_pdev = uncore_extra_pci_dev[pkg].dev[idx];if (filter_pdev) {
Signature · 1 line
2.18.0-rc0
PATCH 22/22 perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support
In reply to Len BrownFrom: Len Brown <[email protected]>
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use the more generic "pmuid" terminology, instead of "pkgid", as the pmu can refer to either packages or die.
--- arch/x86/events/intel/rapl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
arch/x86/events/intel/rapl.c +2 -2
@@ -161,13 +161,13 @@ static u64 rapl_timer_ms;static inline struct rapl_pmu *cpu_to_rapl_pmu (unsigned int cpu) { -unsigned int pkgid =topology_logical_die_id (cpu); +unsigned int pmuid =topology_logical_die_id (cpu);/* * The unsigned check also catches the '-1' return value for non * existent mappings in the topology map. */ -return pkgid < rapl_pmus->maxpkg ? rapl_pmus->pmus[pkgid] :NULL ; +return pmuid < rapl_pmus->maxpkg ? rapl_pmus->pmus[pmuid] :NULL ; }static inline u64 rapl_read_counter (struct perf_event *event)
Signature · 1 line
2.18.0-rc0
PATCH 20/22 hwmon/coretemp: rename internal variables to zones from packages
In reply to Len BrownFrom: Len Brown <[email protected]>
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use the more generic thermal "zone" terminology, instead of "package", as the zones can refer to either packages or die.
--- drivers/hwmon/coretemp.c | 36 ++++++++++++++++++------------------ 1 file changed, 18 insertions(+), 18 deletions(-)
drivers/hwmon/coretemp.c +18 -18
@@ -109,10 +109,10 @@ struct platform_data {struct device_attribute name_attr; }; -/* Keep track of how many package pointers we allocated in init() */ -static int max_packages__read_mostly ; -/* Array of package pointers. Serialized by cpu hotplug lock */ -static struct platform_device **pkg_devices; +/* Keep track of how many zone pointers we allocated in init() */ +static int max_zones__read_mostly ; +/* Array of zone pointers. Serialized by cpu hotplug lock */ +static struct platform_device **zone_devices;static ssize_t show_label (struct device *dev,struct device_attribute *devattr,char *buf)
@@ -435,10 +435,10 @@ static int chk_ucode_version(unsigned int cpu)static struct platform_device *coretemp_get_pdev (unsigned int cpu) { -int pkgid =topology_logical_die_id (cpu); +int id =topology_logical_die_id (cpu); -if (pkgid >=0 && pkgid < max_packages) -return pkg_devices[pkgid]; +if (id >=0 && id < max_zones) +return zone_devices[id];return NULL ; }
@@ -544,7 +544,7 @@ static int coretemp_probe(struct platform_device *pdev)struct device *dev = &pdev->dev;struct platform_data *pdata; -/* Initialize the per-package data structures */ +/* Initialize the per-zone data structures */ pdata =devm_kzalloc (dev,sizeof (struct platform_data ),GFP_KERNEL );if (!pdata)return -ENOMEM ;
@@ -579,13 +579,13 @@ static struct platform_driver coretemp_driver = {static struct platform_device *coretemp_device_add (unsigned int cpu) { -int err, pkgid =topology_logical_die_id (cpu); +int err, zoneid =topology_logical_die_id (cpu);struct platform_device *pdev; -if (pkgid <0 ) +if (zoneid <0 )return ERR_PTR (-ENOMEM ); - pdev =platform_device_alloc (DRVNAME , pkgid); + pdev =platform_device_alloc (DRVNAME , zoneid);if (!pdev)return ERR_PTR (-ENOMEM );
@@ -595,7 +595,7 @@ static struct platform_device *coretemp_device_add(unsigned int cpu)return ERR_PTR (err); } - pkg_devices[pkgid] = pdev; + zone_devices[zoneid] = pdev;return pdev; }
@@ -703,7 +703,7 @@ static int coretemp_cpu_offline(unsigned int cpu)* the rest. */ if (cpumask_empty (&pd->cpumask)) { - pkg_devices[topology_logical_die_id (cpu)] =NULL ; + zone_devices[topology_logical_die_id (cpu)] =NULL ;platform_device_unregister (pdev);return 0 ; }
@@ -741,10 +741,10 @@ static int __init coretemp_init(void)if (!x86_match_cpu (coretemp_ids))return -ENODEV ; - max_packages =topology_max_packages () *topology_max_die_per_package (); - pkg_devices =kcalloc (max_packages,sizeof (struct platform_device *), + max_zones =topology_max_packages () *topology_max_die_per_package (); + zone_devices =kcalloc (max_zones,sizeof (struct platform_device *),GFP_KERNEL ); -if (!pkg_devices) +if (!zone_devices)return -ENOMEM ; err =platform_driver_register (&coretemp_driver);
@@ -760,7 +760,7 @@ static int __init coretemp_init(void) outdrv:platform_driver_unregister (&coretemp_driver); -kfree (pkg_devices); +kfree (zone_devices);return err; }module_init (coretemp_init)
@@ -769,7 +769,7 @@ static void __exit coretemp_exit(void) {cpuhp_remove_state (coretemp_hp_online);platform_driver_unregister (&coretemp_driver); -kfree (pkg_devices); +kfree (zone_devices); }module_exit (coretemp_exit)
Signature · 1 line
2.18.0-rc0
PATCH 19/22 thermal/x86_pkg_temp_thermal: rename internal variables to zones from packages
In reply to Len BrownFrom: Len Brown <[email protected]>
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use the more generic thermal "zone" terminology, instead of "package", as the zones can refer to either packages or die.
--- drivers/thermal/intel/x86_pkg_temp_thermal.c | 142 ++++++++++--------- 1 file changed, 72 insertions(+), 70 deletions(-)
drivers/thermal/intel/x86_pkg_temp_thermal.c +72 -70
@@ -55,7 +55,7 @@ MODULE_PARM_DESC(notify_delay_ms,*/ #define MAX_NUMBER_OF_TRIPS 2 -struct pkg_device { +struct zone_device {int cpu;bool work_scheduled;u32 tj_max;
@@ -70,10 +70,10 @@ static struct thermal_zone_params pkg_temp_tz_params = { .no_hwmon = true, }; -/* Keep track of how many package pointers we allocated in init() */ -static int max_packages__read_mostly ; -/* Array of package pointers */ -static struct pkg_device **packages; +/* Keep track of how many zone pointers we allocated in init() */ +static int max_id__read_mostly ; +/* Array of zone pointers */ +static struct zone_device **zones;/* Serializes interrupt notification, work and hotplug */ static DEFINE_SPINLOCK (pkg_temp_lock);/* Protects zone operation in the work function against hotplug removal */
@@ -120,12 +120,12 @@ static int pkg_temp_debugfs_init(void)* * - Other callsites: Must hold pkg_temp_lock */ -static struct pkg_device *pkg_temp_thermal_get_dev (unsigned int cpu) +static struct zone_device *pkg_temp_thermal_get_dev (unsigned int cpu) { -int pkgid =topology_logical_die_id (cpu); +int id =topology_logical_die_id (cpu); -if (pkgid >=0 && pkgid < max_packages) -return packages[pkgid]; +if (id >=0 && id < max_id) +return zones[id];return NULL ; }
@@ -150,12 +150,13 @@ static int get_tj_max(int cpu, u32 *tj_max)static int sys_get_curr_temp (struct thermal_zone_device *tzd,int *temp) { -struct pkg_device *pkgdev = tzd->devdata; +struct zone_device *zonedev = tzd->devdata;u32 eax, edx; -rdmsr_on_cpu (pkgdev->cpu,MSR_IA32_PACKAGE_THERM_STATUS , &eax, &edx); +rdmsr_on_cpu (zonedev->cpu,MSR_IA32_PACKAGE_THERM_STATUS , + &eax, &edx);if (eax &0x80000000 ) { - *temp = pkgdev->tj_max - ((eax >>16 ) &0x7f ) *1000 ; + *temp = zonedev->tj_max - ((eax >>16 ) &0x7f ) *1000 ;pr_debug ("sys_get_curr_temp %d\n" , *temp);return 0 ; }
@@ -165,7 +166,7 @@ static int sys_get_curr_temp(struct thermal_zone_device *tzd, int *temp)static int sys_get_trip_temp (struct thermal_zone_device *tzd,int trip,int *temp) { -struct pkg_device *pkgdev = tzd->devdata; +struct zone_device *zonedev = tzd->devdata;unsigned long thres_reg_value;u32 mask, shift, eax, edx;int ret;
@@ -181,14 +182,14 @@ static int sys_get_trip_temp(struct thermal_zone_device *tzd, shift =THERM_SHIFT_THRESHOLD0 ; } - ret =rdmsr_on_cpu (pkgdev->cpu,MSR_IA32_PACKAGE_THERM_INTERRUPT , + ret =rdmsr_on_cpu (zonedev->cpu,MSR_IA32_PACKAGE_THERM_INTERRUPT , &eax, &edx);if (ret <0 )return ret; thres_reg_value = (eax & mask) >> shift;if (thres_reg_value) - *temp = pkgdev->tj_max - thres_reg_value *1000 ; + *temp = zonedev->tj_max - thres_reg_value *1000 ;else *temp =0 ;pr_debug ("sys_get_trip_temp %d\n" , *temp);
@@ -199,14 +200,14 @@ static int sys_get_trip_temp(struct thermal_zone_device *tzd,static int sys_set_trip_temp (struct thermal_zone_device *tzd,int trip,int temp) { -struct pkg_device *pkgdev = tzd->devdata; +struct zone_device *zonedev = tzd->devdata;u32 l, h, mask, shift, intr;int ret; -if (trip >=MAX_NUMBER_OF_TRIPS || temp >= pkgdev->tj_max) +if (trip >=MAX_NUMBER_OF_TRIPS || temp >= zonedev->tj_max)return -EINVAL ; - ret =rdmsr_on_cpu (pkgdev->cpu,MSR_IA32_PACKAGE_THERM_INTERRUPT , + ret =rdmsr_on_cpu (zonedev->cpu,MSR_IA32_PACKAGE_THERM_INTERRUPT , &l, &h);if (ret <0 )return ret;
@@ -228,11 +229,12 @@ sys_set_trip_temp(struct thermal_zone_device *tzd, int trip, int temp)if (!temp) { l &= ~intr; }else { - l |= (pkgdev->tj_max - temp)/1000 << shift; + l |= (zonedev->tj_max - temp)/1000 << shift; l |= intr; } -return wrmsr_on_cpu (pkgdev->cpu,MSR_IA32_PACKAGE_THERM_INTERRUPT , l, h); +return wrmsr_on_cpu (zonedev->cpu,MSR_IA32_PACKAGE_THERM_INTERRUPT , + l, h); } static int sys_get_trip_type(struct thermal_zone_device *thermal, int trip,
@@ -287,26 +289,26 @@ static void pkg_temp_thermal_threshold_work_fn(struct work_struct *work) {struct thermal_zone_device *tzone =NULL ;int cpu =smp_processor_id (); -struct pkg_device *pkgdev; +struct zone_device *zonedev;u64 msr_val, wr_val;mutex_lock (&thermal_zone_mutex);spin_lock_irq (&pkg_temp_lock); ++pkg_work_cnt; - pkgdev =pkg_temp_thermal_get_dev (cpu); -if (!pkgdev) { + zonedev =pkg_temp_thermal_get_dev (cpu); +if (!zonedev) {spin_unlock_irq (&pkg_temp_lock);mutex_unlock (&thermal_zone_mutex);return ; } - pkgdev->work_scheduled = false; + zonedev->work_scheduled = false;rdmsrl (MSR_IA32_PACKAGE_THERM_STATUS , msr_val); wr_val = msr_val & ~(THERM_LOG_THRESHOLD0 |THERM_LOG_THRESHOLD1 );if (wr_val != msr_val) {wrmsrl (MSR_IA32_PACKAGE_THERM_STATUS , wr_val); - tzone = pkgdev->tzone; + tzone = zonedev->tzone; }enable_pkg_thres_interrupt ();
@@ -332,7 +334,7 @@ static void pkg_thermal_schedule_work(int cpu, struct delayed_work *work)static int pkg_thermal_notify (u64 msr_val) {int cpu =smp_processor_id (); -struct pkg_device *pkgdev; +struct zone_device *zonedev;unsigned long flags;spin_lock_irqsave (&pkg_temp_lock, flags);
@@ -341,10 +343,10 @@ static int pkg_thermal_notify(u64 msr_val)disable_pkg_thres_interrupt ();/* Work is per package, so scheduling it once is enough. */ - pkgdev =pkg_temp_thermal_get_dev (cpu); -if (pkgdev && !pkgdev->work_scheduled) { - pkgdev->work_scheduled = true; -pkg_thermal_schedule_work (pkgdev->cpu, &pkgdev->work); + zonedev =pkg_temp_thermal_get_dev (cpu); +if (zonedev && !zonedev->work_scheduled) { + zonedev->work_scheduled = true; +pkg_thermal_schedule_work (zonedev->cpu, &zonedev->work); }spin_unlock_irqrestore (&pkg_temp_lock, flags);
@@ -353,12 +355,12 @@ static int pkg_thermal_notify(u64 msr_val)static int pkg_temp_thermal_device_add (unsigned int cpu) { -int pkgid =topology_logical_die_id (cpu); +int id =topology_logical_die_id (cpu);u32 tj_max, eax, ebx, ecx, edx; -struct pkg_device *pkgdev; +struct zone_device *zonedev;int thres_count, err; -if (pkgid >= max_packages) +if (id >= max_id)return -ENOMEM ;cpuid (6 , &eax, &ebx, &ecx, &edx);
@@ -372,51 +374,51 @@ static int pkg_temp_thermal_device_add(unsigned int cpu)if (err)return err; - pkgdev =kzalloc (sizeof (*pkgdev),GFP_KERNEL ); -if (!pkgdev) + zonedev =kzalloc (sizeof (*zonedev),GFP_KERNEL ); +if (!zonedev)return -ENOMEM ; -INIT_DELAYED_WORK (&pkgdev->work, pkg_temp_thermal_threshold_work_fn); - pkgdev->cpu = cpu; - pkgdev->tj_max = tj_max; - pkgdev->tzone =thermal_zone_device_register ("x86_pkg_temp" , +INIT_DELAYED_WORK (&zonedev->work, pkg_temp_thermal_threshold_work_fn); + zonedev->cpu = cpu; + zonedev->tj_max = tj_max; + zonedev->tzone =thermal_zone_device_register ("x86_pkg_temp" , thres_count, (thres_count ==MAX_NUMBER_OF_TRIPS ) ?0x03 :0x01 , - pkgdev, &tzone_ops, &pkg_temp_tz_params,0 ,0 ); -if (IS_ERR (pkgdev->tzone)) { - err =PTR_ERR (pkgdev->tzone); -kfree (pkgdev); + zonedev, &tzone_ops, &pkg_temp_tz_params,0 ,0 ); +if (IS_ERR (zonedev->tzone)) { + err =PTR_ERR (zonedev->tzone); +kfree (zonedev);return err; }/* Store MSR value for package thermal interrupt, to restore at exit */ -rdmsr (MSR_IA32_PACKAGE_THERM_INTERRUPT , pkgdev->msr_pkg_therm_low, - pkgdev->msr_pkg_therm_high); +rdmsr (MSR_IA32_PACKAGE_THERM_INTERRUPT , zonedev->msr_pkg_therm_low, + zonedev->msr_pkg_therm_high); -cpumask_set_cpu (cpu, &pkgdev->cpumask); +cpumask_set_cpu (cpu, &zonedev->cpumask);spin_lock_irq (&pkg_temp_lock); - packages[pkgid] = pkgdev; + zones[id] = zonedev;spin_unlock_irq (&pkg_temp_lock);return 0 ; }static int pkg_thermal_cpu_offline (unsigned int cpu) { -struct pkg_device *pkgdev =pkg_temp_thermal_get_dev (cpu); +struct zone_device *zonedev =pkg_temp_thermal_get_dev (cpu);bool lastcpu, was_target;int target; -if (!pkgdev) +if (!zonedev)return 0 ; - target =cpumask_any_but (&pkgdev->cpumask, cpu); -cpumask_clear_cpu (cpu, &pkgdev->cpumask); + target =cpumask_any_but (&zonedev->cpumask, cpu); +cpumask_clear_cpu (cpu, &zonedev->cpumask); lastcpu = target >= nr_cpu_ids;/* * Remove the sysfs files, if this is the last cpu in the package * before doing further cleanups. */ if (lastcpu) { -struct thermal_zone_device *tzone = pkgdev->tzone; +struct thermal_zone_device *tzone = zonedev->tzone;/* * We must protect against a work function calling
@@ -425,7 +427,7 @@ static int pkg_thermal_cpu_offline(unsigned int cpu)* won't try to call. */ mutex_lock (&thermal_zone_mutex); - pkgdev->tzone =NULL ; + zonedev->tzone =NULL ;mutex_unlock (&thermal_zone_mutex);thermal_zone_device_unregister (tzone);
@@ -439,8 +441,8 @@ static int pkg_thermal_cpu_offline(unsigned int cpu)* one. When we drop the lock, then the interrupt notify function * will see the new target. */ - was_target = pkgdev->cpu == cpu; - pkgdev->cpu = target; + was_target = zonedev->cpu == cpu; + zonedev->cpu = target;/* * If this is the last CPU in the package remove the package
@@ -449,23 +451,23 @@ static int pkg_thermal_cpu_offline(unsigned int cpu)* worker will see the package anymore. */ if (lastcpu) { - packages[topology_logical_die_id (cpu)] =NULL ; + zones[topology_logical_die_id (cpu)] =NULL ;/* After this point nothing touches the MSR anymore. */ wrmsr (MSR_IA32_PACKAGE_THERM_INTERRUPT , - pkgdev->msr_pkg_therm_low, pkgdev->msr_pkg_therm_high); + zonedev->msr_pkg_therm_low, zonedev->msr_pkg_therm_high); }/* * Check whether there is work scheduled and whether the work is * targeted at the outgoing CPU. */ -if (pkgdev->work_scheduled && was_target) { +if (zonedev->work_scheduled && was_target) {/* * To cancel the work we need to drop the lock, otherwise * we might deadlock if the work needs to be flushed. */ spin_unlock_irq (&pkg_temp_lock); -cancel_delayed_work_sync (&pkgdev->work); +cancel_delayed_work_sync (&zonedev->work);spin_lock_irq (&pkg_temp_lock);/* * If this is not the last cpu in the package and the work
@@ -473,21 +475,21 @@ static int pkg_thermal_cpu_offline(unsigned int cpu)* need to reschedule the work, otherwise the interrupt * stays disabled forever. */ -if (!lastcpu && pkgdev->work_scheduled) -pkg_thermal_schedule_work (target, &pkgdev->work); +if (!lastcpu && zonedev->work_scheduled) +pkg_thermal_schedule_work (target, &zonedev->work); }spin_unlock_irq (&pkg_temp_lock);/* Final cleanup if this is the last cpu */ if (lastcpu) -kfree (pkgdev); +kfree (zonedev);return 0 ; }static int pkg_thermal_cpu_online (unsigned int cpu) { -struct pkg_device *pkgdev =pkg_temp_thermal_get_dev (cpu); +struct zone_device *zonedev =pkg_temp_thermal_get_dev (cpu);struct cpuinfo_x86 *c = &cpu_data (cpu);/* Paranoia check */
@@ -495,8 +497,8 @@ static int pkg_thermal_cpu_online(unsigned int cpu)return -ENODEV ;/* If the package exists, nothing to do */ -if (pkgdev) { -cpumask_set_cpu (cpu, &pkgdev->cpumask); +if (zonedev) { +cpumask_set_cpu (cpu, &zonedev->cpumask);return 0 ; }return pkg_temp_thermal_device_add (cpu);
@@ -515,10 +517,10 @@ static int __init pkg_temp_thermal_init(void)if (!x86_match_cpu (pkg_temp_thermal_ids))return -ENODEV ; - max_packages =topology_max_packages () *topology_max_die_per_package (); - packages =kcalloc (max_packages,sizeof (struct pkg_device *), + max_id =topology_max_packages () *topology_max_die_per_package (); + zones =kcalloc (max_id,sizeof (struct zone_device *),GFP_KERNEL ); -if (!packages) +if (!zones)return -ENOMEM ; ret = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "thermal/x86_pkg:online",
@@ -537,7 +539,7 @@ static int __init pkg_temp_thermal_init(void)return 0 ; err: -kfree (packages); +kfree (zones);return ret; }module_init (pkg_temp_thermal_init)
@@ -549,7 +551,7 @@ static void __exit pkg_temp_thermal_exit(void)cpuhp_remove_state (pkg_thermal_hp_state);debugfs_remove_recursive (debugfs); -kfree (packages); +kfree (zones); }module_exit (pkg_temp_thermal_exit)
Signature · 1 line
2.18.0-rc0
PATCH 15/22 topology: Create core_cpus and die_cpus sysfs attributes
In reply to Len BrownFrom: Len Brown <[email protected]>
Create CPU topology sysfs attributes: "core_cpus" and "core_cpus_list"
These attributes represent all of the logical CPUs that share the same core.
These attriutes is synonymous with the existing "thread_siblings" and "thread_siblings_list" attribute, which will be deprecated.
Create CPU topology sysfs attributes: "die_cpus" and "die_cpus_list". These attributes represent all of the logical CPUs that share the same die.
--- Documentation/cputopology.txt | 21 +++++++++++++++------ arch/x86/include/asm/smp.h | 1 + arch/x86/include/asm/topology.h | 1 + arch/x86/kernel/smpboot.c | 22 ++++++++++++++++++++++ arch/x86/xen/smp_pv.c | 1 + drivers/base/topology.c | 12 ++++++++++++ include/linux/topology.h | 3 +++ 7 files changed, 55 insertions(+), 6 deletions(-)
Documentation/cputopology.txt +15 -6
@@ -36,15 +36,15 @@ drawer_id: identifier (rather than the kernel's). The actual value is architecture and platform dependent. -thread_siblings: +core_cpus: - internal kernel map of cpuX's hardware threads within the same - core as cpuX. + internal kernel map of CPUs within the same core. + (deprecated name: "thread_siblings") -thread_siblings_list: +core_cpus_list: - human-readable list of cpuX's hardware threads within the same - core as cpuX. + human-readable list of CPUs within the same core. + (deprecated name: "thread_siblings_list"); package_cpus:
@@ -56,6 +56,14 @@ package_cpus_list: human-readable list of CPUs sharing the same physical_package_id. (deprecated name: "core_siblings_list") +die_cpus: + + internal kernel map of CPUs within the same die. + +die_cpus_list: + + human-readable list of CPUs within the same die. + book_siblings: internal kernel map of cpuX's hardware threads within the same
@@ -93,6 +101,7 @@ these macros in include/asm-XXX/topology.h:: #define topology_drawer_id(cpu) #define topology_sibling_cpumask(cpu) #define topology_core_cpumask(cpu) + #define topology_die_cpumask(cpu) #define topology_book_cpumask(cpu) #define topology_drawer_cpumask(cpu)
arch/x86/include/asm/smp.h +1 -0
@@ -23,6 +23,7 @@ extern unsigned int num_processors;DECLARE_PER_CPU_READ_MOSTLY (cpumask_var_t, cpu_sibling_map);DECLARE_PER_CPU_READ_MOSTLY (cpumask_var_t, cpu_core_map); +DECLARE_PER_CPU_READ_MOSTLY (cpumask_var_t, cpu_die_map);/* cpus sharing the last level cache: */ DECLARE_PER_CPU_READ_MOSTLY (cpumask_var_t, cpu_llc_shared_map);DECLARE_PER_CPU_READ_MOSTLY (u16, cpu_llc_id);
arch/x86/include/asm/topology.h +1 -0
@@ -111,6 +111,7 @@ extern const struct cpumask *cpu_coregroup_mask(int cpu);#define topology_core_id (cpu) (cpu_data(cpu).cpu_core_id)#ifdef CONFIG_SMP +#define topology_die_cpumask (cpu) (per_cpu(cpu_die_map, cpu))#define topology_core_cpumask (cpu) (per_cpu(cpu_core_map, cpu))#define topology_sibling_cpumask (cpu) (per_cpu(cpu_sibling_map, cpu))
arch/x86/kernel/smpboot.c +22 -0
@@ -91,6 +91,10 @@ EXPORT_PER_CPU_SYMBOL(cpu_sibling_map);DEFINE_PER_CPU_READ_MOSTLY (cpumask_var_t, cpu_core_map);EXPORT_PER_CPU_SYMBOL (cpu_core_map); +/* representing HT, core, and die siblings of each logical CPU */ +DEFINE_PER_CPU_READ_MOSTLY (cpumask_var_t, cpu_die_map); +EXPORT_PER_CPU_SYMBOL (cpu_die_map); +DEFINE_PER_CPU_READ_MOSTLY (cpumask_var_t, cpu_llc_shared_map);/* Per CPU bogomips and other parameters */
@@ -509,6 +513,15 @@ static bool match_pkg(struct cpuinfo_x86 *c, struct cpuinfo_x86 *o)return false; } +static bool match_die (struct cpuinfo_x86 *c,struct cpuinfo_x86 *o) +{ +if ((c->phys_proc_id == o->phys_proc_id) && + (c->cpu_die_id == o->cpu_die_id)) +return true; +return false; +} + +#if defined(CONFIG_SCHED_SMT ) || defined(CONFIG_SCHED_MC )static inline int x86_sched_itmt_flags (void ) {
@@ -571,6 +584,7 @@ void set_cpu_sibling_map(int cpu)cpumask_set_cpu (cpu,topology_sibling_cpumask (cpu));cpumask_set_cpu (cpu,cpu_llc_shared_mask (cpu));cpumask_set_cpu (cpu,topology_core_cpumask (cpu)); +cpumask_set_cpu (cpu,topology_die_cpumask (cpu)); c->booted_cores =1 ;return ; }
@@ -619,6 +633,9 @@ void set_cpu_sibling_map(int cpu) }if (match_pkg (c, o) && !topology_same_node (c, o)) x86_has_numa_in_package = true; + +if ((i == cpu) || (has_mp &&match_die (c, o))) +link_mask (topology_die_cpumask, cpu, i); } threads =cpumask_weight (topology_sibling_cpumask (cpu));
@@ -1214,6 +1231,7 @@ static __init void disable_smp(void)physid_set_mask_of_physid (0 , &phys_cpu_present_map);cpumask_set_cpu (0 ,topology_sibling_cpumask (0 ));cpumask_set_cpu (0 ,topology_core_cpumask (0 )); +cpumask_set_cpu (0 ,topology_die_cpumask (0 )); }/*
@@ -1309,6 +1327,7 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus)for_each_possible_cpu (i) {zalloc_cpumask_var (&per_cpu (cpu_sibling_map, i),GFP_KERNEL );zalloc_cpumask_var (&per_cpu (cpu_core_map, i),GFP_KERNEL ); +zalloc_cpumask_var (&per_cpu (cpu_die_map, i),GFP_KERNEL );zalloc_cpumask_var (&per_cpu (cpu_llc_shared_map, i),GFP_KERNEL ); }
@@ -1529,6 +1548,8 @@ static void remove_siblinginfo(int cpu)cpu_data (sibling).booted_cores--; } +for_each_cpu (sibling,topology_die_cpumask (cpu)) +cpumask_clear_cpu (cpu,topology_die_cpumask (sibling));for_each_cpu (sibling,topology_sibling_cpumask (cpu))cpumask_clear_cpu (cpu,topology_sibling_cpumask (sibling));for_each_cpu (sibling,cpu_llc_shared_mask (cpu))
@@ -1536,6 +1557,7 @@ static void remove_siblinginfo(int cpu)cpumask_clear (cpu_llc_shared_mask (cpu));cpumask_clear (topology_sibling_cpumask (cpu));cpumask_clear (topology_core_cpumask (cpu)); +cpumask_clear (topology_die_cpumask (cpu)); c->cpu_core_id =0 ; c->booted_cores =0 ;cpumask_clear_cpu (cpu, cpu_sibling_setup_mask);
arch/x86/xen/smp_pv.c +1 -0
@@ -251,6 +251,7 @@ static void __init xen_pv_smp_prepare_cpus(unsigned int max_cpus)for_each_possible_cpu (i) {zalloc_cpumask_var (&per_cpu (cpu_sibling_map, i),GFP_KERNEL );zalloc_cpumask_var (&per_cpu (cpu_core_map, i),GFP_KERNEL ); +zalloc_cpumask_var (&per_cpu (cpu_die_map, i),GFP_KERNEL );zalloc_cpumask_var (&per_cpu (cpu_llc_shared_map, i),GFP_KERNEL ); }set_cpu_sibling_map (0 );
drivers/base/topology.c +12 -0
@@ -53,10 +53,18 @@ define_siblings_show_func(thread_siblings, sibling_cpumask);static DEVICE_ATTR_RO (thread_siblings);static DEVICE_ATTR_RO (thread_siblings_list); +define_siblings_show_func (core_cpus, sibling_cpumask); +static DEVICE_ATTR_RO (core_cpus); +static DEVICE_ATTR_RO (core_cpus_list); +define_siblings_show_func (core_siblings, core_cpumask);static DEVICE_ATTR_RO (core_siblings);static DEVICE_ATTR_RO (core_siblings_list); +define_siblings_show_func (die_cpus, die_cpumask); +static DEVICE_ATTR_RO (die_cpus); +static DEVICE_ATTR_RO (die_cpus_list); +define_siblings_show_func (package_cpus, core_cpumask);static DEVICE_ATTR_RO (package_cpus);static DEVICE_ATTR_RO (package_cpus_list);
@@ -83,8 +91,12 @@ static struct attribute *default_attrs[] = { &dev_attr_core_id.attr, &dev_attr_thread_siblings.attr, &dev_attr_thread_siblings_list.attr, + &dev_attr_core_cpus.attr, + &dev_attr_core_cpus_list.attr, &dev_attr_core_siblings.attr, &dev_attr_core_siblings_list.attr, + &dev_attr_die_cpus.attr, + &dev_attr_die_cpus_list.attr, &dev_attr_package_cpus.attr, &dev_attr_package_cpus_list.attr, #ifdef CONFIG_SCHED_BOOK
include/linux/topology.h +3 -0
@@ -196,6 +196,9 @@ static inline int cpu_to_mem(int cpu)#ifndef topology_core_cpumask#define topology_core_cpumask (cpu) cpumask_of(cpu)#endif +#ifndef topology_die_cpumask +#define topology_die_cpumask (cpu) cpumask_of(cpu) +#endif #ifdef CONFIG_SCHED_SMT static inline const struct cpumask *cpu_smt_mask (int cpu)
Signature · 1 line
2.18.0-rc0
PATCH 11/22 thermal/x86_pkg_temp_thermal: Support multi-die/package
In reply to Len BrownFrom: Zhang Rui <[email protected]>
Package temperature sensors are actually implemented in hardware per-die. Thus, the new multi-die/package systems sport mulitple package thermal zones for each package.
Update the x86_pkg_temp_thermal to be "multi-die-aware", so it can expose multiple zones per package, instead of just one.
--- drivers/thermal/intel/x86_pkg_temp_thermal.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/thermal/intel/x86_pkg_temp_thermal.c +4 -4
@@ -122,7 +122,7 @@ static int pkg_temp_debugfs_init(void)*/ static struct pkg_device *pkg_temp_thermal_get_dev (unsigned int cpu) { -int pkgid =topology_logical_package_id (cpu); +int pkgid =topology_logical_die_id (cpu);if (pkgid >=0 && pkgid < max_packages)return packages[pkgid];
@@ -353,7 +353,7 @@ static int pkg_thermal_notify(u64 msr_val)static int pkg_temp_thermal_device_add (unsigned int cpu) { -int pkgid =topology_logical_package_id (cpu); +int pkgid =topology_logical_die_id (cpu);u32 tj_max, eax, ebx, ecx, edx;struct pkg_device *pkgdev;int thres_count, err;
@@ -449,7 +449,7 @@ static int pkg_thermal_cpu_offline(unsigned int cpu)* worker will see the package anymore. */ if (lastcpu) { - packages[topology_logical_package_id (cpu)] =NULL ; + packages[topology_logical_die_id (cpu)] =NULL ;/* After this point nothing touches the MSR anymore. */ wrmsr (MSR_IA32_PACKAGE_THERM_INTERRUPT , pkgdev->msr_pkg_therm_low, pkgdev->msr_pkg_therm_high);
@@ -515,7 +515,7 @@ static int __init pkg_temp_thermal_init(void)if (!x86_match_cpu (pkg_temp_thermal_ids))return -ENODEV ; - max_packages =topology_max_packages (); + max_packages =topology_max_packages () *topology_max_die_per_package (); packages =kcalloc (max_packages,sizeof (struct pkg_device *),GFP_KERNEL );if (!packages)
Signature · 1 line
2.18.0-rc0
PATCH 14/22 topology: Create package_cpus sysfs attribute
In reply to Len BrownFrom: Len Brown <[email protected]>
The existing sysfs cpu/topology/core_siblings (and core_siblings_list) attributes are documented, implemented, and used by programs to represent set of logical CPUs sharing the same package.
This makes sense if the next topology level above a core is always a package. But on systems where there is a die topology level between a core and a package, the name and its definition become inconsistent.
So without changing its function, add a name for this map that describes what it actually is -- package CPUs -- the set of CPUs that share the same package.
This new name will be immune to changes in topology, since it describes threads at the current level, not siblings at a contained level.
--- Documentation/cputopology.txt | 12 ++++++------ drivers/base/topology.c | 6 ++++++ 2 files changed, 12 insertions(+), 6 deletions(-)
Documentation/cputopology.txt +6 -6
@@ -46,15 +46,15 @@ thread_siblings_list: human-readable list of cpuX's hardware threads within the same core as cpuX. -core_siblings: +package_cpus: - internal kernel map of cpuX's hardware threads within the same - physical_package_id. + internal kernel map of the CPUs sharing the same physical_package_id. + (deprecated name: "core_siblings") -core_siblings_list: +package_cpus_list: - human-readable list of cpuX's hardware threads within the same - physical_package_id. + human-readable list of CPUs sharing the same physical_package_id. + (deprecated name: "core_siblings_list") book_siblings:
drivers/base/topology.c +6 -0
@@ -57,6 +57,10 @@ define_siblings_show_func(core_siblings, core_cpumask);static DEVICE_ATTR_RO (core_siblings);static DEVICE_ATTR_RO (core_siblings_list); +define_siblings_show_func (package_cpus, core_cpumask); +static DEVICE_ATTR_RO (package_cpus); +static DEVICE_ATTR_RO (package_cpus_list); +#ifdef CONFIG_SCHED_BOOK define_id_show_func (book_id);static DEVICE_ATTR_RO (book_id);
@@ -81,6 +85,8 @@ static struct attribute *default_attrs[] = { &dev_attr_thread_siblings_list.attr, &dev_attr_core_siblings.attr, &dev_attr_core_siblings_list.attr, + &dev_attr_package_cpus.attr, + &dev_attr_package_cpus_list.attr, #ifdef CONFIG_SCHED_BOOK &dev_attr_book_id.attr, &dev_attr_book_siblings.attr,
Signature · 1 line
2.18.0-rc0
PATCH 7/22 x86 topology: Define topology_die_id()
In reply to Len BrownFrom: Len Brown <[email protected]>
topology_die_id(cpu) is a simple macro for use inside the kernel to get the die_id associated with the given cpu.
--- arch/x86/include/asm/topology.h | 1 + 1 file changed, 1 insertion(+)
arch/x86/include/asm/topology.h +1 -0
@@ -106,6 +106,7 @@ extern const struct cpumask *cpu_coregroup_mask(int cpu);#define topology_logical_package_id (cpu) (cpu_data(cpu).logical_proc_id)#define topology_physical_package_id (cpu) (cpu_data(cpu).phys_proc_id) +#define topology_die_id (cpu) (cpu_data(cpu).cpu_die_id)#define topology_core_id (cpu) (cpu_data(cpu).cpu_core_id)#ifdef CONFIG_SMP
Signature · 1 line
2.18.0-rc0
PATCH 5/22 x86 topology: Create topology_max_die_per_package()
In reply to Len BrownFrom: Len Brown <[email protected]>
topology_max_packages() is available to size resources to cover all packages in the system.
But now we have multi-die/package systems, and some resources are per-die.
Create topology_max_die_per_package(), for detecting multi-die/package systems, and sizing any per-die resources.
--- arch/x86/include/asm/processor.h | 1 - arch/x86/include/asm/topology.h | 10 ++++++++++ arch/x86/kernel/cpu/topology.c | 5 ++++- 3 files changed, 14 insertions(+), 2 deletions(-)
arch/x86/include/asm/processor.h +0 -1
@@ -106,7 +106,6 @@ struct cpuinfo_x86 {unsigned long loops_per_jiffy;/* cpuid returned max cores value: */ u16 x86_max_cores; -u16 x86_max_dies;u16 apicid;u16 initial_apicid;u16 x86_clflush_size;
arch/x86/include/asm/topology.h +10 -0
@@ -115,6 +115,13 @@ extern const struct cpumask *cpu_coregroup_mask(int cpu);extern unsigned int __max_logical_packages;#define topology_max_packages () (__max_logical_packages) +extern unsigned int __max_die_per_package; + +static inline int topology_max_die_per_package (void ) +{ +return __max_die_per_package; +} +extern int __max_smt_threads;static inline int topology_max_smt_threads (void )
@@ -131,6 +138,9 @@ bool topology_smt_supported(void);static inline int topology_update_package_map (unsigned int apicid,unsigned int cpu) {return 0 ; }static inline int topology_phys_to_logical_pkg (unsigned int pkg) {return 0 ; } +static inline int topology_phys_to_logical_die (unsigned int die, +unsigned int cpu) {return 0 ; } +static inline int topology_max_die_per_package (void ) {return 1 ; }static inline int topology_max_smt_threads (void ) {return 1 ; }static inline bool topology_is_primary_thread (unsigned int cpu) {return true; }static inline bool topology_smt_supported (void ) {return false; }
arch/x86/kernel/cpu/topology.c +4 -1
@@ -26,6 +26,9 @@#define LEVEL_MAX_SIBLINGS (ebx) ((ebx) & 0xffff)#ifdef CONFIG_SMP +unsigned int __max_die_per_package__read_mostly =1 ; +EXPORT_SYMBOL (__max_die_per_package); +/* * Check if given CPUID extended toplogy "leaf" is implemented */
@@ -146,7 +149,7 @@ int detect_extended_topology(struct cpuinfo_x86 *c) c->apicid = apic->phys_pkg_id (c->initial_apicid,0 ); c->x86_max_cores = (core_level_siblings / smp_num_siblings); - c->x86_max_dies = (die_level_siblings / core_level_siblings); + __max_die_per_package = (die_level_siblings / core_level_siblings);#endif return 0 ; }
Signature · 1 line
2.18.0-rc0
PATCH 2/22 topology: Simplify cputopology.txt formatting and wording
In reply to Len BrownFrom: Len Brown <[email protected]>
Syntax only, no functional or semantic change.
--- Documentation/cputopology.txt | 46 +++++++++++++++++------------------ 1 file changed, 23 insertions(+), 23 deletions(-)
Documentation/cputopology.txt +23 -23
@@ -3,79 +3,79 @@ How CPU topology info is exported via sysfs =========================================== Export CPU topology info via sysfs. Items (attributes) are similar -to /proc/cpuinfo output of some architectures: +to /proc/cpuinfo output of some architectures. They reside in +/sys/devices/system/cpu/cpuX/topology/: -1) /sys/devices/system/cpu/cpuX/topology/physical_package_id: +physical_package_id: physical package id of cpuX. Typically corresponds to a physical socket number, but the actual value is architecture and platform dependent. -2) /sys/devices/system/cpu/cpuX/topology/core_id: +core_id: the CPU core ID of cpuX. Typically it is the hardware platform's identifier (rather than the kernel's). The actual value is architecture and platform dependent. -3) /sys/devices/system/cpu/cpuX/topology/book_id: +book_id: the book ID of cpuX. Typically it is the hardware platform's identifier (rather than the kernel's). The actual value is architecture and platform dependent. -4) /sys/devices/system/cpu/cpuX/topology/drawer_id: +drawer_id: the drawer ID of cpuX. Typically it is the hardware platform's identifier (rather than the kernel's). The actual value is architecture and platform dependent. -5) /sys/devices/system/cpu/cpuX/topology/thread_siblings: +thread_siblings: internal kernel map of cpuX's hardware threads within the same core as cpuX. -6) /sys/devices/system/cpu/cpuX/topology/thread_siblings_list: +thread_siblings_list: human-readable list of cpuX's hardware threads within the same core as cpuX. -7) /sys/devices/system/cpu/cpuX/topology/core_siblings: +core_siblings: internal kernel map of cpuX's hardware threads within the same physical_package_id. -8) /sys/devices/system/cpu/cpuX/topology/core_siblings_list: +core_siblings_list: human-readable list of cpuX's hardware threads within the same physical_package_id. -9) /sys/devices/system/cpu/cpuX/topology/book_siblings: +book_siblings: internal kernel map of cpuX's hardware threads within the same book_id. -10) /sys/devices/system/cpu/cpuX/topology/book_siblings_list: +book_siblings_list: human-readable list of cpuX's hardware threads within the same book_id. -11) /sys/devices/system/cpu/cpuX/topology/drawer_siblings: +drawer_siblings: internal kernel map of cpuX's hardware threads within the same drawer_id. -12) /sys/devices/system/cpu/cpuX/topology/drawer_siblings_list: +drawer_siblings_list: human-readable list of cpuX's hardware threads within the same drawer_id. -To implement it in an architecture-neutral way, a new source file, -drivers/base/topology.c, is to export the 6 to 12 attributes. The book -and drawer related sysfs files will only be created if CONFIG_SCHED_BOOK -and CONFIG_SCHED_DRAWER are selected. +Architecture-neutral, drivers/base/topology.c, exports these attributes. +However, the book and drawer related sysfs files will only be created if +CONFIG_SCHED_BOOK and CONFIG_SCHED_DRAWER are selected, respectively. -CONFIG_SCHED_BOOK and CONFIG_DRAWER are currently only used on s390, where -they reflect the cpu and cache hierarchy. +CONFIG_SCHED_BOOK and CONFIG_SCHED_DRAWER are currently only used on s390, +where they reflect the cpu and cache hierarchy. For an architecture to support this feature, it must define some of these macros in include/asm-XXX/topology.h::
@@ -98,10 +98,10 @@ To be consistent on all architectures, include/linux/topology.h provides default definitions for any of the above macros that are not defined by include/asm-XXX/topology.h: -1) physical_package_id: -1 -2) core_id: 0 -3) sibling_cpumask: just the given CPU -4) core_cpumask: just the given CPU +1) topology_physical_package_id: -1 +2) topology_core_id: 0 +3) topology_sibling_cpumask: just the given CPU +4) topology_core_cpumask: just the given CPU For architectures that don't support books (CONFIG_SCHED_BOOK) there are no default definitions for topology_book_id() and topology_book_cpumask().
Signature · 1 line
2.18.0-rc0
PATCH 3/22 x86 smpboot: Rename match_die() to match_pkg()
In reply to Len BrownFrom: Len Brown <[email protected]>
Syntax only, no functional or semantic change.
This routine matches packages, not die, so name it thus.
--- arch/x86/kernel/smpboot.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
arch/x86/kernel/smpboot.c +3 -3
@@ -455,7 +455,7 @@ static bool match_llc(struct cpuinfo_x86 *c, struct cpuinfo_x86 *o)* multicore group inside a NUMA node. If this happens, we will * discard the MC level of the topology later. */ -static bool match_die (struct cpuinfo_x86 *c,struct cpuinfo_x86 *o) +static bool match_pkg (struct cpuinfo_x86 *c,struct cpuinfo_x86 *o) {if (c->phys_proc_id == o->phys_proc_id)return true;
@@ -546,7 +546,7 @@ void set_cpu_sibling_map(int cpu)for_each_cpu (i, cpu_sibling_setup_mask) { o = &cpu_data (i); -if ((i == cpu) || (has_mp &&match_die (c, o))) { +if ((i == cpu) || (has_mp &&match_pkg (c, o))) {link_mask (topology_core_cpumask, cpu, i);/*
@@ -570,7 +570,7 @@ void set_cpu_sibling_map(int cpu) }else if (i != cpu && !c->booted_cores) c->booted_cores =cpu_data (i).booted_cores; } -if (match_die (c, o) && !topology_same_node (c, o)) +if (match_pkg (c, o) && !topology_same_node (c, o)) x86_has_numa_in_package = true; }
Signature · 1 line
2.18.0-rc0
Peter Zijlstra
maintainerRe: [PATCH 10/22] powercap/intel_rapl: Support multi-die/package
In reply to Len BrownQuoted 32 lines · Len Brown
From: Zhang Rui <[email protected]>
RAPL "package" domains are actually implemented in hardware per-die. Thus, the new multi-die/package systems have mulitple domains within each physical package.
Update the
intel_rapldriver to be "die aware" -- exporting multiple domains within a single package, when present.
No change on single die/package systems.Signed-off-by Zhang Rui <[email protected]>Signed-off-by Len Brown <[email protected]>--- drivers/powercap/intel_rapl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)drivers/powercap/intel_rapl.c +1
-1@@ -266,7 +266,7 @@ static struct rapl_domain *platform_rapl_domain; /* Platform (PSys) domain *//* caller to ensure CPU hotplug lock is held */ static struct rapl_package *rapl_find_package_domain (int cpu) { -int id =topology_physical_package_id (cpu); +int id =topology_logical_die_id (cpu);struct rapl_package *rp;list_for_each_entry (rp, &rapl_packages, plist) {
drivers/powercap/intel_rapl.c static void rapl_remove_package(struct rapl_package *rp) · line 1464
Quoted 7 lines
drivers/powercap/intel_rapl.c +1
-1@@ -1459,7 +1459,7 @@ static void rapl_remove_package(struct rapl_package *rp)/* called from CPU hotplug notifier, hotplug lock held */ static struct rapl_package *rapl_add_package (int cpu) { -int id =topology_physical_package_id (cpu); +int id =topology_logical_die_id (cpu);struct rapl_package *rp;int ret;
Both functions are still misnomers. rapl_find_package_domain() does in fact now do rapl_find_die_domain(), right? Same for rapl_add_package()
Peter Zijlstra
maintainerRe: [PATCH 16/22] perf/x86/intel/uncore: Support multi-die/package
In reply to Len Brownarch/x86/events/intel/uncore.c static int __init intel_uncore_init(void) · line 1416
Quoted 6 lines · Len Brown
arch/x86/events/intel/uncore.c +1
-1@@ -1411,7 +1413,7 @@ static int __init intel_uncore_init(void)if (boot_cpu_has (X86_FEATURE_HYPERVISOR ))return -ENODEV ; - max_packages =topology_max_packages (); + max_packages =topology_max_packages () *topology_max_die_per_package ();
That's max_dies..
Quoted 5 lines
arch/x86/events/intel/uncore.c +0
-0uncore_init = (struct intel_uncore_init_fun *)id->driver_data;if (uncore_init->pci_init) {Signature · 1 line
2.18.0-rc0
Peter Zijlstra
maintainerRe: [PATCH 16/22] perf/x86/intel/uncore: Support multi-die/package
In reply to Len Brownarch/x86/events/intel/uncore.c static int uncore_event_cpu_online(unsigned int cpu) · line 1231
Quoted 9 lines · Len Brown
arch/x86/events/intel/uncore.c +1
-1@@ -1223,7 +1225,7 @@ static int uncore_event_cpu_online(unsigned int cpu)struct intel_uncore_box *box;int i, ret, pkg, target; - pkg =topology_logical_package_id (cpu); + pkg =topology_logical_die_id (cpu); ret =allocate_boxes (types, pkg, cpu);if (ret)return ret;
'pkg' != 'die'
Re: [PATCH 10/22] powercap/intel_rapl: Support multi-die/package
In reply to Peter ZijlstraQuoted 20 lines · Peter Zijlstra
Quoted 18 lines · Len Brown
From: Zhang Rui <[email protected]>
RAPL "package" domains are actually implemented in hardware per-die. Thus, the new multi-die/package systems have mulitple domains within each physical package.
Update the
intel_rapldriver to be "die aware" -- exporting multiple domains within a single package, when present.
No change on single die/package systems.Signed-off-by Zhang Rui <[email protected]>Signed-off-by Len Brown <[email protected]>--- drivers/powercap/intel_rapl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
Quoted 14 lines
drivers/powercap/intel_rapl.c static struct rapl_domain *platform_rapl_domain; /* Platform (PSys) domain */ · line 270
Quoted 11 lines
drivers/powercap/intel_rapl.c +1
-1@@ -266,7 +266,7 @@ static struct rapl_domain *platform_rapl_domain; /* Platform (PSys) domain *//* caller to ensure CPU hotplug lock is held */ static struct rapl_package *rapl_find_package_domain (int cpu) { -int id =topology_physical_package_id (cpu); +int id =topology_logical_die_id (cpu);struct rapl_package *rp;Both functions are still misnomers.
rapl_find_package_domain()does in fact now dorapl_find_die_domain(), right? Same forrapl_add_package()
A "RAPL Package Domain" (rapl_package, above) is a known proper noun -- it is a known documented capability.
When there could be just 1 die in a package, the name of this capability also always matched the scope of a physical package.
Now that some products have two die in the same package, there can be two of these inside a package, and they are associated with each die.
There are no plans to re-name the Package RAPL Domain capability in the hardware documentation.
Similarly, there are no plans to re-name any of the other "PACKAGE" scoped MSRs to have "DIE" in their name instead. The ship with those names sailed long ago.
I think the code above reflects its function, and that somebody maintaining it will be clear on this. That is important, because in the future, there will be a concept of PACKAGE scoped MSRs that span multiple DIE...
cheers,
Len Brown, Intel Open Source Technology Center
Re: [PATCH 16/22] perf/x86/intel/uncore: Support multi-die/package
In reply to Peter ZijlstraQuoted 10 lines · Peter Zijlstra
arch/x86/events/intel/uncore.c static int __init intel_uncore_init(void) · line 1416
Quoted 6 lines · Len Brown
arch/x86/events/intel/uncore.c +1
-1@@ -1411,7 +1413,7 @@ static int __init intel_uncore_init(void)if (boot_cpu_has (X86_FEATURE_HYPERVISOR ))return -ENODEV ; - max_packages =topology_max_packages (); + max_packages =topology_max_packages () *topology_max_die_per_package ();That's
max_dies..
To keep the patch with this functional change minimal and obvious, the rename of this variable was segregated into patch 22, which is re-names only.
Len Brown, Intel Open Source Technology Center
Re: [PATCH 16/22] perf/x86/intel/uncore: Support multi-die/package
In reply to Peter ZijlstraQuoted 6 lines · Peter Zijlstra
Quoted 4 lines · Len Brown
arch/x86/events/intel/uncore.c +0
-0@@ -1223,7 +1225,7 @@ static int uncore_event_cpu_online(unsigned int cpu)struct intel_uncore_box *box;int i, ret, pkg, target;
Quoted 7 lines
arch/x86/events/intel/uncore.c static int uncore_event_cpu_online(unsigned int cpu) · line 1231
Quoted 5 lines
arch/x86/events/intel/uncore.c +1
-1- pkg =topology_logical_package_id (cpu); + pkg =topology_logical_die_id (cpu); ret =allocate_boxes (types, pkg, cpu);if (ret)return ret;'pkg' != 'die'
To keep the patch with this functional change minimal and obvious, the rename of this variable was segregated into patch 22, which is re-names only.
Len Brown, Intel Open Source Technology Center
Re: [PATCH 21/22] perf/x86/intel/uncore: renames in response to multi-die/pkg support
In reply to Len BrownQuoted 8 lines · Len Brown
From: Len Brown <[email protected]>
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use the more generic "box" terminology, instead of "pkg", as the boxes can refer to either packages or die.
I think the "box" terminology in perf uncore has different meaning. It stands for an uncore PMU unit on a socket/die. I think it may be better use "die" to replace the "pkg".
How about the patch as below?
Thanks,
Kan
From 94187403be80b1c1ac13929973958c046a952278 Mon Sep 17 00:00:00 2001
From: Kan Liang <[email protected]>
Date: Thu, 9 May 2019 07:19:21 -0700
Subject: [PATCH 1/2] perf/x86/intel/uncore: Renames in response to multi-die/pkg support
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use "die" terminology, instead of "pkg".
For previous platforms which doesn't have multi-die, "die" is identical as "pkg".
--- arch/x86/events/intel/uncore.c | 74 ++++++++++++++++++------------------ arch/x86/events/intel/uncore.h | 5 ++- arch/x86/events/intel/uncore_snbep.c | 2 +- 3 files changed, 41 insertions(+), 40 deletions(-)
arch/x86/events/intel/uncore.c +1 -1
@@ -14,7 +14,7 @@ struct pci_driver *uncore_pci_driver;DEFINE_RAW_SPINLOCK (pci2phy_map_lock);struct list_head pci2phy_map_head =LIST_HEAD_INIT (pci2phy_map_head);struct pci_extra_dev *uncore_extra_pci_dev; -static int max_packages; +static int max_dies;/* mask of cpus that collect uncore events */ static cpumask_t uncore_cpu_mask;
@@ -100,13 +100,13 @@ ssize_t uncore_event_show(struct kobject *kobj, struct intel_uncore_box *uncore_pmu_to_box(struct intel_uncore_pmu
*pmu, int cpu)
{
- unsigned int pkgid = topology_logical_die_id(cpu);
+ unsigned int dieid = topology_logical_die_id(cpu);
/*
* The unsigned check also catches the '-1' return value for non
* existent mappings in the topology map.
*/
- return pkgid < max_packages ? pmu->boxes[pkgid] : NULL;
+ return dieid < max_dies ? pmu->boxes[dieid] : NULL;
}
u64 uncore_msr_read_counter(struct intel_uncore_box *box, structperf_event *event)
@@ -311,7 +311,7 @@ static struct intel_uncore_box *uncore_alloc_box(struct intel_uncore_type *type,
uncore_pmu_init_hrtimer(box);
box->cpu = -1;
box->pci_phys_id = -1;
- box->pkgid = -1;
+ box->dieid = -1;
/* set default hrtimer timeout */
box->hrtimer_duration = UNCORE_PMU_HRTIMER_INTERVAL;
@@ -826,10 +826,10 @@ static void uncore_pmu_unregister(struct intel_uncore_pmu *pmu)
static void uncore_free_boxes(struct intel_uncore_pmu *pmu) {
- int pkg;
+ int die;
- for (pkg = 0; pkg < max_packages; pkg++)
- kfree(pmu->boxes[pkg]);
+ for (die = 0; die < max_dies; die++)
+ kfree(pmu->boxes[die]);
kfree(pmu->boxes);
}
@@ -866,7 +866,7 @@ static int __init uncore_type_init(struct intel_uncore_type *type, bool setid)
if (!pmus)
return -ENOMEM;
- size = max_packages * sizeof(struct intel_uncore_box *);
+ size = max_dies * sizeof(struct intel_uncore_box *);
for (i = 0; i < type->num_boxes; i++) {
pmus[i].func_id = setid ? i : -1;
@@ -936,21 +936,21 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id
struct intel_uncore_type *type;
struct intel_uncore_pmu *pmu = NULL;
struct intel_uncore_box *box;
- int phys_id, pkg, ret;
+ int phys_id, die, ret;
phys_id = uncore_pcibus_to_physid(pdev->bus); if (phys_id < 0)
return -ENODEV;
- pkg = (topology_max_die_per_package() > 1) ? phys_id :
+ die = (topology_max_die_per_package() > 1) ? phys_id :
topology_phys_to_logical_pkg(phys_id);
- if (pkg < 0)
+ if (die < 0)
return -EINVAL;
if (UNCORE_PCI_DEV_TYPE(id->driver_data) == UNCORE_EXTRA_PCI_DEV) {
int idx = UNCORE_PCI_DEV_IDX(id->driver_data);
- uncore_extra_pci_dev[pkg].dev[idx] = pdev;
+ uncore_extra_pci_dev[die].dev[idx] = pdev;
pci_set_drvdata(pdev, NULL);
return 0;
}
@@ -989,7 +989,7 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id
pmu = &type->pmus[UNCORE_PCI_DEV_IDX(id->driver_data)];
}
- if (WARN_ON_ONCE(pmu->boxes[pkg] != NULL))
+ if (WARN_ON_ONCE(pmu->boxes[die] != NULL))
return -EINVAL;
box = uncore_alloc_box(type, NUMA_NO_NODE);
@@ -1003,13 +1003,13 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id
atomic_inc(&box->refcnt);
box->pci_phys_id = phys_id;
- box->pkgid = pkg;
+ box->dieid = die;
box->pci_dev = pdev;
box->pmu = pmu;
uncore_box_init(box);
pci_set_drvdata(pdev, box);
- pmu->boxes[pkg] = box;
+ pmu->boxes[die] = box;
if (atomic_inc_return(&pmu->activeboxes) > 1)
return 0;
@@ -1017,7 +1017,7 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id
ret = uncore_pmu_register(pmu);
if (ret) {
pci_set_drvdata(pdev, NULL);
- pmu->boxes[pkg] = NULL;
+ pmu->boxes[die] = NULL;
uncore_box_exit(box);
kfree(box);
}
@@ -1028,17 +1028,17 @@ static void uncore_pci_remove(struct pci_dev *pdev)
{
struct intel_uncore_box *box;
struct intel_uncore_pmu *pmu;
- int i, phys_id, pkg;
+ int i, phys_id, die;
phys_id = uncore_pcibus_to_physid(pdev->bus);
box = pci_get_drvdata(pdev);
if (!box) {
- pkg = (topology_max_die_per_package() > 1) ? phys_id :
+ die = (topology_max_die_per_package() > 1) ? phys_id :
topology_phys_to_logical_pkg(phys_id);
for (i = 0; i < UNCORE_EXTRA_PCI_DEV_MAX; i++) {
- if (uncore_extra_pci_dev[pkg].dev[i] == pdev) {
- uncore_extra_pci_dev[pkg].dev[i] = NULL;
+ if (uncore_extra_pci_dev[die].dev[i] == pdev) {
+ uncore_extra_pci_dev[die].dev[i] = NULL;
break;
}
}
@@ -1051,7 +1051,7 @@ static void uncore_pci_remove(struct pci_dev *pdev)
return;
pci_set_drvdata(pdev, NULL);
- pmu->boxes[box->pkgid] = NULL;
+ pmu->boxes[box->dieid] = NULL;
if (atomic_dec_return(&pmu->activeboxes) == 0)
uncore_pmu_unregister(pmu);
uncore_box_exit(box);
@@ -1063,7 +1063,7 @@ static int __init uncore_pci_init(void)
size_t size;
int ret;
- size = max_packages * sizeof(struct pci_extra_dev);
+ size = max_dies * sizeof(struct pci_extra_dev);
uncore_extra_pci_dev = kzalloc(size, GFP_KERNEL);
if (!uncore_extra_pci_dev) {
ret = -ENOMEM;
@@ -1110,11 +1110,11 @@ static void uncore_change_type_ctx(struct intel_uncore_type *type, int old_cpu,
{
struct intel_uncore_pmu *pmu = type->pmus;
struct intel_uncore_box *box;
- int i, pkg;
+ int i, die;
- pkg = topology_logical_die_id(old_cpu < 0 ? new_cpu : old_cpu);
+ die = topology_logical_die_id(old_cpu < 0 ? new_cpu : old_cpu);
for (i = 0; i < type->num_boxes; i++, pmu++) {
- box = pmu->boxes[pkg];
+ box = pmu->boxes[die];
if (!box)
continue;
@@ -1147,7 +1147,7 @@ static int uncore_event_cpu_offline(unsigned int cpu)
struct intel_uncore_type *type, **types = uncore_msr_uncores; struct intel_uncore_pmu *pmu;
struct intel_uncore_box *box;
- int i, pkg, target;
+ int i, die, target;
/* Check if exiting cpu is used for collecting uncore events */ if (!cpumask_test_and_clear_cpu(cpu, &uncore_cpu_mask))
@@ -1166,12 +1166,12 @@ static int uncore_event_cpu_offline(unsigned int cpu)
unref:
/* Clear the references */
- pkg = topology_logical_die_id(cpu);
+ die = topology_logical_die_id(cpu);
for (; *types; types++) {
type = *types;
pmu = type->pmus;
for (i = 0; i < type->num_boxes; i++, pmu++) {
- box = pmu->boxes[pkg];
+ box = pmu->boxes[die];
if (box && atomic_dec_return(&box->refcnt) == 0)
uncore_box_exit(box);
}
@@ -1180,7 +1180,7 @@ static int uncore_event_cpu_offline(unsigned int cpu)
}
static int allocate_boxes(struct intel_uncore_type **types,
- unsigned int pkg, unsigned int cpu)
+ unsigned int die, unsigned int cpu)
{
struct intel_uncore_box *box, *tmp;
struct intel_uncore_type *type;
@@ -1193,20 +1193,20 @@ static int allocate_boxes(struct intel_uncore_type **types,
type = *types;
pmu = type->pmus;
for (i = 0; i < type->num_boxes; i++, pmu++) {
- if (pmu->boxes[pkg])
+ if (pmu->boxes[die])
continue;
box = uncore_alloc_box(type, cpu_to_node(cpu)); if (!box)
goto cleanup;
box->pmu = pmu;
- box->pkgid = pkg;
+ box->dieid = die;
list_add(&box->active_list, &allocated);
}
}
/* Install them in the pmus */
list_for_each_entry_safe(box, tmp, &allocated, active_list) {
list_del_init(&box->active_list);
- box->pmu->boxes[pkg] = box;
+ box->pmu->boxes[die] = box;
}
return 0;
@@ -1223,10 +1223,10 @@ static int uncore_event_cpu_online(unsigned int cpu)
struct intel_uncore_type *type, **types = uncore_msr_uncores; struct intel_uncore_pmu *pmu;
struct intel_uncore_box *box;
- int i, ret, pkg, target;
+ int i, ret, die, target;
- pkg = topology_logical_die_id(cpu);
- ret = allocate_boxes(types, pkg, cpu);
+ die = topology_logical_die_id(cpu);
+ ret = allocate_boxes(types, die, cpu);
if (ret)
return ret;
@@ -1234,7 +1234,7 @@ static int uncore_event_cpu_online(unsigned int cpu)
type = *types;
pmu = type->pmus;
for (i = 0; i < type->num_boxes; i++, pmu++) {
- box = pmu->boxes[pkg];
+ box = pmu->boxes[die];
if (box && atomic_inc_return(&box->refcnt) == 1)
uncore_box_init(box);
}
@@ -1413,7 +1413,7 @@ static int __init intel_uncore_init(void)
if (boot_cpu_has(X86_FEATURE_HYPERVISOR))
return -ENODEV;
- max_packages = topology_max_packages() * topology_max_die_per_package();
+ max_dies = topology_max_packages() * topology_max_die_per_package();
uncore_init = (struct intel_uncore_init_fun *)id->driver_data; if (uncore_init->pci_init) {
arch/x86/events/intel/uncore.h +3 -2
@@ -108,7 +108,8 @@ struct intel_uncore_extra_reg {struct intel_uncore_box {int pci_phys_id; -int pkgid;/* Logical package ID */ +/* Logical die ID, which is identical as pkg id for single die system. */ +int dieid;int n_active;/* number of active events */ int n_events;int cpu;/* cpu to collect events */
@@ -467,7 +468,7 @@ static inline void uncore_box_exit(struct intel_uncore_box *box)static inline bool uncore_box_is_fake (struct intel_uncore_box *box) { -return (box->pkgid <0 ); +return (box->dieid <0 ); } static inline struct intel_uncore_pmu *uncore_event_to_pmu(struct
perf_event *event)
arch/x86/events/intel/uncore_snbep.c +1 -1
@@ -1058,7 +1058,7 @@ static void snbep_qpi_enable_event(struct intel_uncore_box *box, struct perf_eveif (reg1->idx !=EXTRA_REG_NONE ) {int idx = box->pmu->pmu_idx +SNBEP_PCI_QPI_PORT0_FILTER ; -int pkg = box->pkgid; +int pkg = box->dieid;struct pci_dev *filter_pdev = uncore_extra_pci_dev[pkg].dev[idx];if (filter_pdev) {
Signature · 1 line
2.7.4
Re: [PATCH 22/22] perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support
In reply to Len BrownQuoted 7 lines · Len Brown
From: Len Brown <[email protected]>
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use the more generic "pmuid" terminology, instead of "pkgid", as the pmu can refer to either packages or die.
The perf rapl "pmu" in the code is cross the pkg/die. We only register one rapl pmu for whole system for now. I think it may be better use "die" to replace the "pkg" as well.
How about the patch as below?
Thanks,
Kan
From a898867b24ed8ea3582b0c8b3218838701249065 Mon Sep 17 00:00:00 2001
From: Kan Liang <[email protected]>
Date: Thu, 9 May 2019 07:21:35 -0700
Subject: [PATCH 2/2] perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use "die" terminology, instead of "pkg".
For previous platforms which doesn't have multi-die, "die" is identical as "pkg".
--- arch/x86/events/intel/rapl.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
arch/x86/events/intel/rapl.c +7 -7
@@ -148,7 +148,7 @@ struct rapl_pmu {struct rapl_pmus {struct pmu pmu; -unsigned int maxpkg; +unsigned int maxdie;struct rapl_pmu *pmus[]; };
@@ -161,13 +161,13 @@ static u64 rapl_timer_ms;static inline struct rapl_pmu *cpu_to_rapl_pmu (unsigned int cpu) { -unsigned int pkgid =topology_logical_die_id (cpu); +unsigned int dieid =topology_logical_die_id (cpu);/* * The unsigned check also catches the '-1' return value for non * existent mappings in the topology map. */ -return pkgid < rapl_pmus->maxpkg ? rapl_pmus->pmus[pkgid] :NULL ; +return dieid < rapl_pmus->maxdie ? rapl_pmus->pmus[dieid] :NULL ; }static inline u64 rapl_read_counter (struct perf_event *event)
@@ -668,22 +668,22 @@ static void cleanup_rapl_pmus(void) {int i; -for (i =0 ; i < rapl_pmus->maxpkg; i++) +for (i =0 ; i < rapl_pmus->maxdie; i++)kfree (rapl_pmus->pmus[i]);kfree (rapl_pmus); }static int __init init_rapl_pmus (void ) { -int maxpkg =topology_max_packages () *topology_max_die_per_package (); +int maxdie =topology_max_packages () *topology_max_die_per_package ();size_t size; - size =sizeof (*rapl_pmus) + maxpkg *sizeof (struct rapl_pmu *); + size =sizeof (*rapl_pmus) + maxdie *sizeof (struct rapl_pmu *); rapl_pmus =kzalloc (size,GFP_KERNEL );if (!rapl_pmus)return -ENOMEM ; - rapl_pmus->maxpkg = maxpkg; + rapl_pmus->maxdie = maxdie; rapl_pmus->pmu.attr_groups = rapl_attr_groups; rapl_pmus->pmu.task_ctx_nr = perf_invalid_context; rapl_pmus->pmu.event_init = rapl_pmu_event_init;
Signature · 1 line
2.7.4
Re: [PATCH 22/22] perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support
Quoted 4 lines · Kan
The perf rapl "pmu" in the code is cross the pkg/die. We only register one rapl pmu for whole system for now.
I think it may be better use "die" to replace the "pkg" as well. How about the patch as below?
Fine with me!
I've replaced my patch with yours in the series.
thanks,
Len Brown, Intel Open Source Technology Center
Re: [PATCH 21/22] perf/x86/intel/uncore: renames in response to multi-die/pkg support
In reply to Liang, KanQuoted 5 lines · Kan
I think the "box" terminology in perf uncore has different meaning. It stands for an uncore PMU unit on a socket/die.
I think it may be better use "die" to replace the "pkg". How about the patch as below?
Also fine with me.
And I've replaced my rename patch with yours here too.
thanks,
Len Brown, Intel Open Source Technology Center