summaryrefslogtreecommitdiff
path: root/arch/x86/events
diff options
context:
space:
mode:
authorAndré Fabian Silva Delgado <emulatorman@parabola.nu>2016-10-20 00:10:27 -0300
committerAndré Fabian Silva Delgado <emulatorman@parabola.nu>2016-10-20 00:10:27 -0300
commitd0b2f91bede3bd5e3d24dd6803e56eee959c1797 (patch)
tree7fee4ab0509879c373c4f2cbd5b8a5be5b4041ee /arch/x86/events
parente914f8eb445e8f74b00303c19c2ffceaedd16a05 (diff)
Linux-libre 4.8.2-gnupck-4.8.2-gnu
Diffstat (limited to 'arch/x86/events')
-rw-r--r--arch/x86/events/amd/core.c6
-rw-r--r--arch/x86/events/amd/ibs.c75
-rw-r--r--arch/x86/events/amd/iommu.c2
-rw-r--r--arch/x86/events/amd/power.c60
-rw-r--r--arch/x86/events/amd/uncore.c144
-rw-r--r--arch/x86/events/core.c140
-rw-r--r--arch/x86/events/intel/bts.c126
-rw-r--r--arch/x86/events/intel/core.c220
-rw-r--r--arch/x86/events/intel/cqm.c49
-rw-r--r--arch/x86/events/intel/cstate.c98
-rw-r--r--arch/x86/events/intel/lbr.c124
-rw-r--r--arch/x86/events/intel/rapl.c116
-rw-r--r--arch/x86/events/intel/uncore.c223
-rw-r--r--arch/x86/events/intel/uncore.h6
-rw-r--r--arch/x86/events/intel/uncore_snb.c81
-rw-r--r--arch/x86/events/intel/uncore_snbep.c96
-rw-r--r--arch/x86/events/msr.c63
-rw-r--r--arch/x86/events/perf_event.h12
18 files changed, 977 insertions, 664 deletions
diff --git a/arch/x86/events/amd/core.c b/arch/x86/events/amd/core.c
index a8309ea67..f5f4b3fbb 100644
--- a/arch/x86/events/amd/core.c
+++ b/arch/x86/events/amd/core.c
@@ -370,13 +370,13 @@ static int amd_pmu_cpu_prepare(int cpu)
WARN_ON_ONCE(cpuc->amd_nb);
if (!x86_pmu.amd_nb_constraints)
- return NOTIFY_OK;
+ return 0;
cpuc->amd_nb = amd_alloc_nb(cpu);
if (!cpuc->amd_nb)
- return NOTIFY_BAD;
+ return -ENOMEM;
- return NOTIFY_OK;
+ return 0;
}
static void amd_pmu_cpu_starting(int cpu)
diff --git a/arch/x86/events/amd/ibs.c b/arch/x86/events/amd/ibs.c
index feb90f673..b26ee32f7 100644
--- a/arch/x86/events/amd/ibs.c
+++ b/arch/x86/events/amd/ibs.c
@@ -7,7 +7,8 @@
*/
#include <linux/perf_event.h>
-#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/export.h>
#include <linux/pci.h>
#include <linux/ptrace.h>
#include <linux/syscore_ops.h>
@@ -655,8 +656,12 @@ fail:
}
if (event->attr.sample_type & PERF_SAMPLE_RAW) {
- raw.size = sizeof(u32) + ibs_data.size;
- raw.data = ibs_data.data;
+ raw = (struct perf_raw_record){
+ .frag = {
+ .size = sizeof(u32) + ibs_data.size,
+ .data = ibs_data.data,
+ },
+ };
data.raw = &raw;
}
@@ -721,13 +726,10 @@ static __init int perf_ibs_pmu_init(struct perf_ibs *perf_ibs, char *name)
return ret;
}
-static __init int perf_event_ibs_init(void)
+static __init void perf_event_ibs_init(void)
{
struct attribute **attr = ibs_op_format_attrs;
- if (!ibs_caps)
- return -ENODEV; /* ibs not supported by the cpu */
-
perf_ibs_pmu_init(&perf_ibs_fetch, "ibs_fetch");
if (ibs_caps & IBS_CAPS_OPCNT) {
@@ -738,13 +740,11 @@ static __init int perf_event_ibs_init(void)
register_nmi_handler(NMI_LOCAL, perf_ibs_nmi_handler, 0, "perf_ibs");
pr_info("perf: AMD IBS detected (0x%08x)\n", ibs_caps);
-
- return 0;
}
#else /* defined(CONFIG_PERF_EVENTS) && defined(CONFIG_CPU_SUP_AMD) */
-static __init int perf_event_ibs_init(void) { return 0; }
+static __init void perf_event_ibs_init(void) { }
#endif
@@ -921,7 +921,7 @@ static inline int get_ibs_lvt_offset(void)
return val & IBSCTL_LVT_OFFSET_MASK;
}
-static void setup_APIC_ibs(void *dummy)
+static void setup_APIC_ibs(void)
{
int offset;
@@ -936,7 +936,7 @@ failed:
smp_processor_id());
}
-static void clear_APIC_ibs(void *dummy)
+static void clear_APIC_ibs(void)
{
int offset;
@@ -945,18 +945,24 @@ static void clear_APIC_ibs(void *dummy)
setup_APIC_eilvt(offset, 0, APIC_EILVT_MSG_FIX, 1);
}
+static int x86_pmu_amd_ibs_starting_cpu(unsigned int cpu)
+{
+ setup_APIC_ibs();
+ return 0;
+}
+
#ifdef CONFIG_PM
static int perf_ibs_suspend(void)
{
- clear_APIC_ibs(NULL);
+ clear_APIC_ibs();
return 0;
}
static void perf_ibs_resume(void)
{
ibs_eilvt_setup();
- setup_APIC_ibs(NULL);
+ setup_APIC_ibs();
}
static struct syscore_ops perf_ibs_syscore_ops = {
@@ -975,27 +981,15 @@ static inline void perf_ibs_pm_init(void) { }
#endif
-static int
-perf_ibs_cpu_notifier(struct notifier_block *self, unsigned long action, void *hcpu)
+static int x86_pmu_amd_ibs_dying_cpu(unsigned int cpu)
{
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_STARTING:
- setup_APIC_ibs(NULL);
- break;
- case CPU_DYING:
- clear_APIC_ibs(NULL);
- break;
- default:
- break;
- }
-
- return NOTIFY_OK;
+ clear_APIC_ibs();
+ return 0;
}
static __init int amd_ibs_init(void)
{
u32 caps;
- int ret = -EINVAL;
caps = __get_ibs_caps();
if (!caps)
@@ -1004,22 +998,25 @@ static __init int amd_ibs_init(void)
ibs_eilvt_setup();
if (!ibs_eilvt_valid())
- goto out;
+ return -EINVAL;
perf_ibs_pm_init();
- cpu_notifier_register_begin();
+
ibs_caps = caps;
/* make ibs_caps visible to other cpus: */
smp_mb();
- smp_call_function(setup_APIC_ibs, NULL, 1);
- __perf_cpu_notifier(perf_ibs_cpu_notifier);
- cpu_notifier_register_done();
+ /*
+ * x86_pmu_amd_ibs_starting_cpu will be called from core on
+ * all online cpus.
+ */
+ cpuhp_setup_state(CPUHP_AP_PERF_X86_AMD_IBS_STARTING,
+ "AP_PERF_X86_AMD_IBS_STARTING",
+ x86_pmu_amd_ibs_starting_cpu,
+ x86_pmu_amd_ibs_dying_cpu);
- ret = perf_event_ibs_init();
-out:
- if (ret)
- pr_err("Failed to setup IBS, %d\n", ret);
- return ret;
+ perf_event_ibs_init();
+
+ return 0;
}
/* Since we need the pci subsystem to init ibs we can't do this earlier: */
diff --git a/arch/x86/events/amd/iommu.c b/arch/x86/events/amd/iommu.c
index 6011a573d..b28200dea 100644
--- a/arch/x86/events/amd/iommu.c
+++ b/arch/x86/events/amd/iommu.c
@@ -12,7 +12,7 @@
*/
#include <linux/perf_event.h>
-#include <linux/module.h>
+#include <linux/init.h>
#include <linux/cpumask.h>
#include <linux/slab.h>
diff --git a/arch/x86/events/amd/power.c b/arch/x86/events/amd/power.c
index 55a3529db..9842270ed 100644
--- a/arch/x86/events/amd/power.c
+++ b/arch/x86/events/amd/power.c
@@ -228,12 +228,12 @@ static struct pmu pmu_class = {
.read = pmu_event_read,
};
-static void power_cpu_exit(int cpu)
+static int power_cpu_exit(unsigned int cpu)
{
int target;
if (!cpumask_test_and_clear_cpu(cpu, &cpu_mask))
- return;
+ return 0;
/*
* Find a new CPU on the same compute unit, if was set in cpumask
@@ -245,9 +245,10 @@ static void power_cpu_exit(int cpu)
cpumask_set_cpu(target, &cpu_mask);
perf_pmu_migrate_context(&pmu_class, cpu, target);
}
+ return 0;
}
-static void power_cpu_init(int cpu)
+static int power_cpu_init(unsigned int cpu)
{
int target;
@@ -255,7 +256,7 @@ static void power_cpu_init(int cpu)
* 1) If any CPU is set at cpu_mask in the same compute unit, do
* nothing.
* 2) If no CPU is set at cpu_mask in the same compute unit,
- * set current STARTING CPU.
+ * set current ONLINE CPU.
*
* Note: if there is a CPU aside of the new one already in the
* sibling mask, then it is also in cpu_mask.
@@ -263,33 +264,9 @@ static void power_cpu_init(int cpu)
target = cpumask_any_but(topology_sibling_cpumask(cpu), cpu);
if (target >= nr_cpumask_bits)
cpumask_set_cpu(cpu, &cpu_mask);
+ return 0;
}
-static int
-power_cpu_notifier(struct notifier_block *self, unsigned long action, void *hcpu)
-{
- unsigned int cpu = (long)hcpu;
-
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_DOWN_FAILED:
- case CPU_STARTING:
- power_cpu_init(cpu);
- break;
- case CPU_DOWN_PREPARE:
- power_cpu_exit(cpu);
- break;
- default:
- break;
- }
-
- return NOTIFY_OK;
-}
-
-static struct notifier_block power_cpu_notifier_nb = {
- .notifier_call = power_cpu_notifier,
- .priority = CPU_PRI_PERF,
-};
-
static const struct x86_cpu_id cpu_match[] = {
{ .vendor = X86_VENDOR_AMD, .family = 0x15 },
{},
@@ -297,7 +274,7 @@ static const struct x86_cpu_id cpu_match[] = {
static int __init amd_power_pmu_init(void)
{
- int cpu, target, ret;
+ int ret;
if (!x86_match_cpu(cpu_match))
return 0;
@@ -312,38 +289,25 @@ static int __init amd_power_pmu_init(void)
return -ENODEV;
}
- cpu_notifier_register_begin();
- /* Choose one online core of each compute unit. */
- for_each_online_cpu(cpu) {
- target = cpumask_first(topology_sibling_cpumask(cpu));
- if (!cpumask_test_cpu(target, &cpu_mask))
- cpumask_set_cpu(target, &cpu_mask);
- }
+ cpuhp_setup_state(CPUHP_AP_PERF_X86_AMD_POWER_ONLINE,
+ "AP_PERF_X86_AMD_POWER_ONLINE",
+ power_cpu_init, power_cpu_exit);
ret = perf_pmu_register(&pmu_class, "power", -1);
if (WARN_ON(ret)) {
pr_warn("AMD Power PMU registration failed\n");
- goto out;
+ return ret;
}
- __register_cpu_notifier(&power_cpu_notifier_nb);
-
pr_info("AMD Power PMU detected\n");
-
-out:
- cpu_notifier_register_done();
-
return ret;
}
module_init(amd_power_pmu_init);
static void __exit amd_power_pmu_exit(void)
{
- cpu_notifier_register_begin();
- __unregister_cpu_notifier(&power_cpu_notifier_nb);
- cpu_notifier_register_done();
-
+ cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_AMD_POWER_ONLINE);
perf_pmu_unregister(&pmu_class);
}
module_exit(amd_power_pmu_exit);
diff --git a/arch/x86/events/amd/uncore.c b/arch/x86/events/amd/uncore.c
index 98ac57381..65577f081 100644
--- a/arch/x86/events/amd/uncore.c
+++ b/arch/x86/events/amd/uncore.c
@@ -29,6 +29,8 @@
#define COUNTER_SHIFT 16
+static HLIST_HEAD(uncore_unused_list);
+
struct amd_uncore {
int id;
int refcnt;
@@ -39,7 +41,7 @@ struct amd_uncore {
cpumask_t *active_mask;
struct pmu *pmu;
struct perf_event *events[MAX_COUNTERS];
- struct amd_uncore *free_when_cpu_online;
+ struct hlist_node node;
};
static struct amd_uncore * __percpu *amd_uncore_nb;
@@ -306,6 +308,7 @@ static int amd_uncore_cpu_up_prepare(unsigned int cpu)
uncore_nb->msr_base = MSR_F15H_NB_PERF_CTL;
uncore_nb->active_mask = &amd_nb_active_mask;
uncore_nb->pmu = &amd_nb_pmu;
+ uncore_nb->id = -1;
*per_cpu_ptr(amd_uncore_nb, cpu) = uncore_nb;
}
@@ -319,6 +322,7 @@ static int amd_uncore_cpu_up_prepare(unsigned int cpu)
uncore_l2->msr_base = MSR_F16H_L2I_PERF_CTL;
uncore_l2->active_mask = &amd_l2_active_mask;
uncore_l2->pmu = &amd_l2_pmu;
+ uncore_l2->id = -1;
*per_cpu_ptr(amd_uncore_l2, cpu) = uncore_l2;
}
@@ -348,7 +352,7 @@ amd_uncore_find_online_sibling(struct amd_uncore *this,
continue;
if (this->id == that->id) {
- that->free_when_cpu_online = this;
+ hlist_add_head(&this->node, &uncore_unused_list);
this = that;
break;
}
@@ -358,7 +362,7 @@ amd_uncore_find_online_sibling(struct amd_uncore *this,
return this;
}
-static void amd_uncore_cpu_starting(unsigned int cpu)
+static int amd_uncore_cpu_starting(unsigned int cpu)
{
unsigned int eax, ebx, ecx, edx;
struct amd_uncore *uncore;
@@ -384,6 +388,19 @@ static void amd_uncore_cpu_starting(unsigned int cpu)
uncore = amd_uncore_find_online_sibling(uncore, amd_uncore_l2);
*per_cpu_ptr(amd_uncore_l2, cpu) = uncore;
}
+
+ return 0;
+}
+
+static void uncore_clean_online(void)
+{
+ struct amd_uncore *uncore;
+ struct hlist_node *n;
+
+ hlist_for_each_entry_safe(uncore, n, &uncore_unused_list, node) {
+ hlist_del(&uncore->node);
+ kfree(uncore);
+ }
}
static void uncore_online(unsigned int cpu,
@@ -391,20 +408,21 @@ static void uncore_online(unsigned int cpu,
{
struct amd_uncore *uncore = *per_cpu_ptr(uncores, cpu);
- kfree(uncore->free_when_cpu_online);
- uncore->free_when_cpu_online = NULL;
+ uncore_clean_online();
if (cpu == uncore->cpu)
cpumask_set_cpu(cpu, uncore->active_mask);
}
-static void amd_uncore_cpu_online(unsigned int cpu)
+static int amd_uncore_cpu_online(unsigned int cpu)
{
if (amd_uncore_nb)
uncore_online(cpu, amd_uncore_nb);
if (amd_uncore_l2)
uncore_online(cpu, amd_uncore_l2);
+
+ return 0;
}
static void uncore_down_prepare(unsigned int cpu,
@@ -433,13 +451,15 @@ static void uncore_down_prepare(unsigned int cpu,
}
}
-static void amd_uncore_cpu_down_prepare(unsigned int cpu)
+static int amd_uncore_cpu_down_prepare(unsigned int cpu)
{
if (amd_uncore_nb)
uncore_down_prepare(cpu, amd_uncore_nb);
if (amd_uncore_l2)
uncore_down_prepare(cpu, amd_uncore_l2);
+
+ return 0;
}
static void uncore_dead(unsigned int cpu, struct amd_uncore * __percpu *uncores)
@@ -454,74 +474,19 @@ static void uncore_dead(unsigned int cpu, struct amd_uncore * __percpu *uncores)
*per_cpu_ptr(uncores, cpu) = NULL;
}
-static void amd_uncore_cpu_dead(unsigned int cpu)
+static int amd_uncore_cpu_dead(unsigned int cpu)
{
if (amd_uncore_nb)
uncore_dead(cpu, amd_uncore_nb);
if (amd_uncore_l2)
uncore_dead(cpu, amd_uncore_l2);
-}
-
-static int
-amd_uncore_cpu_notifier(struct notifier_block *self, unsigned long action,
- void *hcpu)
-{
- unsigned int cpu = (long)hcpu;
-
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_UP_PREPARE:
- if (amd_uncore_cpu_up_prepare(cpu))
- return notifier_from_errno(-ENOMEM);
- break;
-
- case CPU_STARTING:
- amd_uncore_cpu_starting(cpu);
- break;
-
- case CPU_ONLINE:
- amd_uncore_cpu_online(cpu);
- break;
-
- case CPU_DOWN_PREPARE:
- amd_uncore_cpu_down_prepare(cpu);
- break;
-
- case CPU_UP_CANCELED:
- case CPU_DEAD:
- amd_uncore_cpu_dead(cpu);
- break;
-
- default:
- break;
- }
-
- return NOTIFY_OK;
-}
-
-static struct notifier_block amd_uncore_cpu_notifier_block = {
- .notifier_call = amd_uncore_cpu_notifier,
- .priority = CPU_PRI_PERF + 1,
-};
-
-static void __init init_cpu_already_online(void *dummy)
-{
- unsigned int cpu = smp_processor_id();
-
- amd_uncore_cpu_starting(cpu);
- amd_uncore_cpu_online(cpu);
-}
-static void cleanup_cpu_online(void *dummy)
-{
- unsigned int cpu = smp_processor_id();
-
- amd_uncore_cpu_dead(cpu);
+ return 0;
}
static int __init amd_uncore_init(void)
{
- unsigned int cpu, cpu2;
int ret = -ENODEV;
if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD)
@@ -558,38 +523,29 @@ static int __init amd_uncore_init(void)
ret = 0;
}
- if (ret)
- goto fail_nodev;
-
- cpu_notifier_register_begin();
-
- /* init cpus already online before registering for hotplug notifier */
- for_each_online_cpu(cpu) {
- ret = amd_uncore_cpu_up_prepare(cpu);
- if (ret)
- goto fail_online;
- smp_call_function_single(cpu, init_cpu_already_online, NULL, 1);
- }
-
- __register_cpu_notifier(&amd_uncore_cpu_notifier_block);
- cpu_notifier_register_done();
-
+ /*
+ * Install callbacks. Core will call them for each online cpu.
+ */
+ if (cpuhp_setup_state(CPUHP_PERF_X86_AMD_UNCORE_PREP,
+ "PERF_X86_AMD_UNCORE_PREP",
+ amd_uncore_cpu_up_prepare, amd_uncore_cpu_dead))
+ goto fail_l2;
+
+ if (cpuhp_setup_state(CPUHP_AP_PERF_X86_AMD_UNCORE_STARTING,
+ "AP_PERF_X86_AMD_UNCORE_STARTING",
+ amd_uncore_cpu_starting, NULL))
+ goto fail_prep;
+ if (cpuhp_setup_state(CPUHP_AP_PERF_X86_AMD_UNCORE_ONLINE,
+ "AP_PERF_X86_AMD_UNCORE_ONLINE",
+ amd_uncore_cpu_online,
+ amd_uncore_cpu_down_prepare))
+ goto fail_start;
return 0;
-
-fail_online:
- for_each_online_cpu(cpu2) {
- if (cpu2 == cpu)
- break;
- smp_call_function_single(cpu, cleanup_cpu_online, NULL, 1);
- }
- cpu_notifier_register_done();
-
- /* amd_uncore_nb/l2 should have been freed by cleanup_cpu_online */
- amd_uncore_nb = amd_uncore_l2 = NULL;
-
- if (boot_cpu_has(X86_FEATURE_PERFCTR_L2))
- perf_pmu_unregister(&amd_l2_pmu);
+fail_start:
+ cpuhp_remove_state(CPUHP_AP_PERF_X86_AMD_UNCORE_STARTING);
+fail_prep:
+ cpuhp_remove_state(CPUHP_PERF_X86_AMD_UNCORE_PREP);
fail_l2:
if (boot_cpu_has(X86_FEATURE_PERFCTR_NB))
perf_pmu_unregister(&amd_nb_pmu);
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c
index 91eac3962..d0efb5cb1 100644
--- a/arch/x86/events/core.c
+++ b/arch/x86/events/core.c
@@ -17,7 +17,8 @@
#include <linux/notifier.h>
#include <linux/hardirq.h>
#include <linux/kprobes.h>
-#include <linux/module.h>
+#include <linux/export.h>
+#include <linux/init.h>
#include <linux/kdebug.h>
#include <linux/sched.h>
#include <linux/uaccess.h>
@@ -262,10 +263,13 @@ static bool check_hw_exists(void)
return true;
msr_fail:
- pr_cont("Broken PMU hardware detected, using software events only.\n");
- printk("%sFailed to access perfctr msr (MSR %x is %Lx)\n",
- boot_cpu_has(X86_FEATURE_HYPERVISOR) ? KERN_INFO : KERN_ERR,
- reg, val_new);
+ if (boot_cpu_has(X86_FEATURE_HYPERVISOR)) {
+ pr_cont("PMU not available due to virtualization, using software events only.\n");
+ } else {
+ pr_cont("Broken PMU hardware detected, using software events only.\n");
+ pr_err("Failed to access perfctr msr (MSR %x is %Lx)\n",
+ reg, val_new);
+ }
return false;
}
@@ -1477,49 +1481,49 @@ NOKPROBE_SYMBOL(perf_event_nmi_handler);
struct event_constraint emptyconstraint;
struct event_constraint unconstrained;
-static int
-x86_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu)
+static int x86_pmu_prepare_cpu(unsigned int cpu)
{
- unsigned int cpu = (long)hcpu;
struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu);
- int i, ret = NOTIFY_OK;
-
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_UP_PREPARE:
- for (i = 0 ; i < X86_PERF_KFREE_MAX; i++)
- cpuc->kfree_on_online[i] = NULL;
- if (x86_pmu.cpu_prepare)
- ret = x86_pmu.cpu_prepare(cpu);
- break;
-
- case CPU_STARTING:
- if (x86_pmu.cpu_starting)
- x86_pmu.cpu_starting(cpu);
- break;
+ int i;
- case CPU_ONLINE:
- for (i = 0 ; i < X86_PERF_KFREE_MAX; i++) {
- kfree(cpuc->kfree_on_online[i]);
- cpuc->kfree_on_online[i] = NULL;
- }
- break;
+ for (i = 0 ; i < X86_PERF_KFREE_MAX; i++)
+ cpuc->kfree_on_online[i] = NULL;
+ if (x86_pmu.cpu_prepare)
+ return x86_pmu.cpu_prepare(cpu);
+ return 0;
+}
- case CPU_DYING:
- if (x86_pmu.cpu_dying)
- x86_pmu.cpu_dying(cpu);
- break;
+static int x86_pmu_dead_cpu(unsigned int cpu)
+{
+ if (x86_pmu.cpu_dead)
+ x86_pmu.cpu_dead(cpu);
+ return 0;
+}
- case CPU_UP_CANCELED:
- case CPU_DEAD:
- if (x86_pmu.cpu_dead)
- x86_pmu.cpu_dead(cpu);
- break;
+static int x86_pmu_online_cpu(unsigned int cpu)
+{
+ struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu);
+ int i;
- default:
- break;
+ for (i = 0 ; i < X86_PERF_KFREE_MAX; i++) {
+ kfree(cpuc->kfree_on_online[i]);
+ cpuc->kfree_on_online[i] = NULL;
}
+ return 0;
+}
- return ret;
+static int x86_pmu_starting_cpu(unsigned int cpu)
+{
+ if (x86_pmu.cpu_starting)
+ x86_pmu.cpu_starting(cpu);
+ return 0;
+}
+
+static int x86_pmu_dying_cpu(unsigned int cpu)
+{
+ if (x86_pmu.cpu_dying)
+ x86_pmu.cpu_dying(cpu);
+ return 0;
}
static void __init pmu_check_apic(void)
@@ -1622,6 +1626,29 @@ ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr, cha
}
EXPORT_SYMBOL_GPL(events_sysfs_show);
+ssize_t events_ht_sysfs_show(struct device *dev, struct device_attribute *attr,
+ char *page)
+{
+ struct perf_pmu_events_ht_attr *pmu_attr =
+ container_of(attr, struct perf_pmu_events_ht_attr, attr);
+
+ /*
+ * Report conditional events depending on Hyper-Threading.
+ *
+ * This is overly conservative as usually the HT special
+ * handling is not needed if the other CPU thread is idle.
+ *
+ * Note this does not (and cannot) handle the case when thread
+ * siblings are invisible, for example with virtualization
+ * if they are owned by some other guest. The user tool
+ * has to re-read when a thread sibling gets onlined later.
+ */
+ return sprintf(page, "%s",
+ topology_max_smt_threads() > 1 ?
+ pmu_attr->event_str_ht :
+ pmu_attr->event_str_noht);
+}
+
EVENT_ATTR(cpu-cycles, CPU_CYCLES );
EVENT_ATTR(instructions, INSTRUCTIONS );
EVENT_ATTR(cache-references, CACHE_REFERENCES );
@@ -1764,10 +1791,39 @@ static int __init init_hw_perf_events(void)
pr_info("... fixed-purpose events: %d\n", x86_pmu.num_counters_fixed);
pr_info("... event mask: %016Lx\n", x86_pmu.intel_ctrl);
- perf_pmu_register(&pmu, "cpu", PERF_TYPE_RAW);
- perf_cpu_notifier(x86_pmu_notifier);
+ /*
+ * Install callbacks. Core will call them for each online
+ * cpu.
+ */
+ err = cpuhp_setup_state(CPUHP_PERF_X86_PREPARE, "PERF_X86_PREPARE",
+ x86_pmu_prepare_cpu, x86_pmu_dead_cpu);
+ if (err)
+ return err;
+
+ err = cpuhp_setup_state(CPUHP_AP_PERF_X86_STARTING,
+ "AP_PERF_X86_STARTING", x86_pmu_starting_cpu,
+ x86_pmu_dying_cpu);
+ if (err)
+ goto out;
+
+ err = cpuhp_setup_state(CPUHP_AP_PERF_X86_ONLINE, "AP_PERF_X86_ONLINE",
+ x86_pmu_online_cpu, NULL);
+ if (err)
+ goto out1;
+
+ err = perf_pmu_register(&pmu, "cpu", PERF_TYPE_RAW);
+ if (err)
+ goto out2;
return 0;
+
+out2:
+ cpuhp_remove_state(CPUHP_AP_PERF_X86_ONLINE);
+out1:
+ cpuhp_remove_state(CPUHP_AP_PERF_X86_STARTING);
+out:
+ cpuhp_remove_state(CPUHP_PERF_X86_PREPARE);
+ return err;
}
early_initcall(init_hw_perf_events);
diff --git a/arch/x86/events/intel/bts.c b/arch/x86/events/intel/bts.c
index 0a6e393a2..982c9e31d 100644
--- a/arch/x86/events/intel/bts.c
+++ b/arch/x86/events/intel/bts.c
@@ -31,7 +31,17 @@
struct bts_ctx {
struct perf_output_handle handle;
struct debug_store ds_back;
- int started;
+ int state;
+};
+
+/* BTS context states: */
+enum {
+ /* no ongoing AUX transactions */
+ BTS_STATE_STOPPED = 0,
+ /* AUX transaction is on, BTS tracing is disabled */
+ BTS_STATE_INACTIVE,
+ /* AUX transaction is on, BTS tracing is running */
+ BTS_STATE_ACTIVE,
};
static DEFINE_PER_CPU(struct bts_ctx, bts_ctx);
@@ -204,6 +214,15 @@ static void bts_update(struct bts_ctx *bts)
static int
bts_buffer_reset(struct bts_buffer *buf, struct perf_output_handle *handle);
+/*
+ * Ordering PMU callbacks wrt themselves and the PMI is done by means
+ * of bts::state, which:
+ * - is set when bts::handle::event is valid, that is, between
+ * perf_aux_output_begin() and perf_aux_output_end();
+ * - is zero otherwise;
+ * - is ordered against bts::handle::event with a compiler barrier.
+ */
+
static void __bts_event_start(struct perf_event *event)
{
struct bts_ctx *bts = this_cpu_ptr(&bts_ctx);
@@ -221,10 +240,13 @@ static void __bts_event_start(struct perf_event *event)
/*
* local barrier to make sure that ds configuration made it
- * before we enable BTS
+ * before we enable BTS and bts::state goes ACTIVE
*/
wmb();
+ /* INACTIVE/STOPPED -> ACTIVE */
+ WRITE_ONCE(bts->state, BTS_STATE_ACTIVE);
+
intel_pmu_enable_bts(config);
}
@@ -251,9 +273,6 @@ static void bts_event_start(struct perf_event *event, int flags)
__bts_event_start(event);
- /* PMI handler: this counter is running and likely generating PMIs */
- ACCESS_ONCE(bts->started) = 1;
-
return;
fail_end_stop:
@@ -263,30 +282,34 @@ fail_stop:
event->hw.state = PERF_HES_STOPPED;
}
-static void __bts_event_stop(struct perf_event *event)
+static void __bts_event_stop(struct perf_event *event, int state)
{
+ struct bts_ctx *bts = this_cpu_ptr(&bts_ctx);
+
+ /* ACTIVE -> INACTIVE(PMI)/STOPPED(->stop()) */
+ WRITE_ONCE(bts->state, state);
+
/*
* No extra synchronization is mandated by the documentation to have
* BTS data stores globally visible.
*/
intel_pmu_disable_bts();
-
- if (event->hw.state & PERF_HES_STOPPED)
- return;
-
- ACCESS_ONCE(event->hw.state) |= PERF_HES_STOPPED;
}
static void bts_event_stop(struct perf_event *event, int flags)
{
struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events);
struct bts_ctx *bts = this_cpu_ptr(&bts_ctx);
- struct bts_buffer *buf = perf_get_aux(&bts->handle);
+ struct bts_buffer *buf = NULL;
+ int state = READ_ONCE(bts->state);
- /* PMI handler: don't restart this counter */
- ACCESS_ONCE(bts->started) = 0;
+ if (state == BTS_STATE_ACTIVE)
+ __bts_event_stop(event, BTS_STATE_STOPPED);
- __bts_event_stop(event);
+ if (state != BTS_STATE_STOPPED)
+ buf = perf_get_aux(&bts->handle);
+
+ event->hw.state |= PERF_HES_STOPPED;
if (flags & PERF_EF_UPDATE) {
bts_update(bts);
@@ -296,6 +319,7 @@ static void bts_event_stop(struct perf_event *event, int flags)
bts->handle.head =
local_xchg(&buf->data_size,
buf->nr_pages << PAGE_SHIFT);
+
perf_aux_output_end(&bts->handle, local_xchg(&buf->data_size, 0),
!!local_xchg(&buf->lost, 0));
}
@@ -310,8 +334,20 @@ static void bts_event_stop(struct perf_event *event, int flags)
void intel_bts_enable_local(void)
{
struct bts_ctx *bts = this_cpu_ptr(&bts_ctx);
+ int state = READ_ONCE(bts->state);
+
+ /*
+ * Here we transition from INACTIVE to ACTIVE;
+ * if we instead are STOPPED from the interrupt handler,
+ * stay that way. Can't be ACTIVE here though.
+ */
+ if (WARN_ON_ONCE(state == BTS_STATE_ACTIVE))
+ return;
+
+ if (state == BTS_STATE_STOPPED)
+ return;
- if (bts->handle.event && bts->started)
+ if (bts->handle.event)
__bts_event_start(bts->handle.event);
}
@@ -319,8 +355,15 @@ void intel_bts_disable_local(void)
{
struct bts_ctx *bts = this_cpu_ptr(&bts_ctx);
+ /*
+ * Here we transition from ACTIVE to INACTIVE;
+ * do nothing for STOPPED or INACTIVE.
+ */
+ if (READ_ONCE(bts->state) != BTS_STATE_ACTIVE)
+ return;
+
if (bts->handle.event)
- __bts_event_stop(bts->handle.event);
+ __bts_event_stop(bts->handle.event, BTS_STATE_INACTIVE);
}
static int
@@ -335,8 +378,6 @@ bts_buffer_reset(struct bts_buffer *buf, struct perf_output_handle *handle)
return 0;
head = handle->head & ((buf->nr_pages << PAGE_SHIFT) - 1);
- if (WARN_ON_ONCE(head != local_read(&buf->head)))
- return -EINVAL;
phys = &buf->buf[buf->cur_buf];
space = phys->offset + phys->displacement + phys->size - head;
@@ -403,22 +444,37 @@ bts_buffer_reset(struct bts_buffer *buf, struct perf_output_handle *handle)
int intel_bts_interrupt(void)
{
+ struct debug_store *ds = this_cpu_ptr(&cpu_hw_events)->ds;
struct bts_ctx *bts = this_cpu_ptr(&bts_ctx);
struct perf_event *event = bts->handle.event;
struct bts_buffer *buf;
s64 old_head;
- int err;
+ int err = -ENOSPC, handled = 0;
- if (!event || !bts->started)
- return 0;
+ /*
+ * The only surefire way of knowing if this NMI is ours is by checking
+ * the write ptr against the PMI threshold.
+ */
+ if (ds && (ds->bts_index >= ds->bts_interrupt_threshold))
+ handled = 1;
+
+ /*
+ * this is wrapped in intel_bts_enable_local/intel_bts_disable_local,
+ * so we can only be INACTIVE or STOPPED
+ */
+ if (READ_ONCE(bts->state) == BTS_STATE_STOPPED)
+ return handled;
buf = perf_get_aux(&bts->handle);
+ if (!buf)
+ return handled;
+
/*
* Skip snapshot counters: they don't use the interrupt, but
* there's no other way of telling, because the pointer will
* keep moving
*/
- if (!buf || buf->snapshot)
+ if (buf->snapshot)
return 0;
old_head = local_read(&buf->head);
@@ -426,18 +482,27 @@ int intel_bts_interrupt(void)
/* no new data */
if (old_head == local_read(&buf->head))
- return 0;
+ return handled;
perf_aux_output_end(&bts->handle, local_xchg(&buf->data_size, 0),
!!local_xchg(&buf->lost, 0));
buf = perf_aux_output_begin(&bts->handle, event);
- if (!buf)
- return 1;
+ if (buf)
+ err = bts_buffer_reset(buf, &bts->handle);
+
+ if (err) {
+ WRITE_ONCE(bts->state, BTS_STATE_STOPPED);
- err = bts_buffer_reset(buf, &bts->handle);
- if (err)
- perf_aux_output_end(&bts->handle, 0, false);
+ if (buf) {
+ /*
+ * BTS_STATE_STOPPED should be visible before
+ * cleared handle::event
+ */
+ barrier();
+ perf_aux_output_end(&bts->handle, 0, false);
+ }
+ }
return 1;
}
@@ -519,7 +584,8 @@ static __init int bts_init(void)
if (!boot_cpu_has(X86_FEATURE_DTES64) || !x86_pmu.bts)
return -ENODEV;
- bts_pmu.capabilities = PERF_PMU_CAP_AUX_NO_SG | PERF_PMU_CAP_ITRACE;
+ bts_pmu.capabilities = PERF_PMU_CAP_AUX_NO_SG | PERF_PMU_CAP_ITRACE |
+ PERF_PMU_CAP_EXCLUSIVE;
bts_pmu.task_ctx_nr = perf_sw_context;
bts_pmu.event_init = bts_event_init;
bts_pmu.add = bts_event_add;
diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c
index 9b4f9d3ce..4c9a79b9c 100644
--- a/arch/x86/events/intel/core.c
+++ b/arch/x86/events/intel/core.c
@@ -16,6 +16,7 @@
#include <asm/cpufeature.h>
#include <asm/hardirq.h>
+#include <asm/intel-family.h>
#include <asm/apic.h>
#include "../perf_event.h"
@@ -185,7 +186,7 @@ static struct event_constraint intel_slm_event_constraints[] __read_mostly =
EVENT_CONSTRAINT_END
};
-struct event_constraint intel_skl_event_constraints[] = {
+static struct event_constraint intel_skl_event_constraints[] = {
FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */
FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */
FIXED_EVENT_CONSTRAINT(0x0300, 2), /* CPU_CLK_UNHALTED.REF */
@@ -204,10 +205,8 @@ struct event_constraint intel_skl_event_constraints[] = {
};
static struct extra_reg intel_knl_extra_regs[] __read_mostly = {
- INTEL_UEVENT_EXTRA_REG(0x01b7,
- MSR_OFFCORE_RSP_0, 0x7f9ffbffffull, RSP_0),
- INTEL_UEVENT_EXTRA_REG(0x02b7,
- MSR_OFFCORE_RSP_1, 0x3f9ffbffffull, RSP_1),
+ INTEL_UEVENT_EXTRA_REG(0x01b7, MSR_OFFCORE_RSP_0, 0x799ffbb6e7ull, RSP_0),
+ INTEL_UEVENT_EXTRA_REG(0x02b7, MSR_OFFCORE_RSP_1, 0x399ffbffe7ull, RSP_1),
EVENT_EXTRA_END
};
@@ -243,14 +242,51 @@ EVENT_ATTR_STR(mem-loads, mem_ld_nhm, "event=0x0b,umask=0x10,ldlat=3");
EVENT_ATTR_STR(mem-loads, mem_ld_snb, "event=0xcd,umask=0x1,ldlat=3");
EVENT_ATTR_STR(mem-stores, mem_st_snb, "event=0xcd,umask=0x2");
-struct attribute *nhm_events_attrs[] = {
+static struct attribute *nhm_events_attrs[] = {
EVENT_PTR(mem_ld_nhm),
NULL,
};
-struct attribute *snb_events_attrs[] = {
+/*
+ * topdown events for Intel Core CPUs.
+ *
+ * The events are all in slots, which is a free slot in a 4 wide
+ * pipeline. Some events are already reported in slots, for cycle
+ * events we multiply by the pipeline width (4).
+ *
+ * With Hyper Threading on, topdown metrics are either summed or averaged
+ * between the threads of a core: (count_t0 + count_t1).
+ *
+ * For the average case the metric is always scaled to pipeline width,
+ * so we use factor 2 ((count_t0 + count_t1) / 2 * 4)
+ */
+
+EVENT_ATTR_STR_HT(topdown-total-slots, td_total_slots,
+ "event=0x3c,umask=0x0", /* cpu_clk_unhalted.thread */
+ "event=0x3c,umask=0x0,any=1"); /* cpu_clk_unhalted.thread_any */
+EVENT_ATTR_STR_HT(topdown-total-slots.scale, td_total_slots_scale, "4", "2");
+EVENT_ATTR_STR(topdown-slots-issued, td_slots_issued,
+ "event=0xe,umask=0x1"); /* uops_issued.any */
+EVENT_ATTR_STR(topdown-slots-retired, td_slots_retired,
+ "event=0xc2,umask=0x2"); /* uops_retired.retire_slots */
+EVENT_ATTR_STR(topdown-fetch-bubbles, td_fetch_bubbles,
+ "event=0x9c,umask=0x1"); /* idq_uops_not_delivered_core */
+EVENT_ATTR_STR_HT(topdown-recovery-bubbles, td_recovery_bubbles,
+ "event=0xd,umask=0x3,cmask=1", /* int_misc.recovery_cycles */
+ "event=0xd,umask=0x3,cmask=1,any=1"); /* int_misc.recovery_cycles_any */
+EVENT_ATTR_STR_HT(topdown-recovery-bubbles.scale, td_recovery_bubbles_scale,
+ "4", "2");
+
+static struct attribute *snb_events_attrs[] = {
EVENT_PTR(mem_ld_snb),
EVENT_PTR(mem_st_snb),
+ EVENT_PTR(td_slots_issued),
+ EVENT_PTR(td_slots_retired),
+ EVENT_PTR(td_fetch_bubbles),
+ EVENT_PTR(td_total_slots),
+ EVENT_PTR(td_total_slots_scale),
+ EVENT_PTR(td_recovery_bubbles),
+ EVENT_PTR(td_recovery_bubbles_scale),
NULL,
};
@@ -280,7 +316,7 @@ static struct event_constraint intel_hsw_event_constraints[] = {
EVENT_CONSTRAINT_END
};
-struct event_constraint intel_bdw_event_constraints[] = {
+static struct event_constraint intel_bdw_event_constraints[] = {
FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */
FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */
FIXED_EVENT_CONSTRAINT(0x0300, 2), /* CPU_CLK_UNHALTED.REF */
@@ -1361,6 +1397,29 @@ static __initconst const u64 atom_hw_cache_event_ids
},
};
+EVENT_ATTR_STR(topdown-total-slots, td_total_slots_slm, "event=0x3c");
+EVENT_ATTR_STR(topdown-total-slots.scale, td_total_slots_scale_slm, "2");
+/* no_alloc_cycles.not_delivered */
+EVENT_ATTR_STR(topdown-fetch-bubbles, td_fetch_bubbles_slm,
+ "event=0xca,umask=0x50");
+EVENT_ATTR_STR(topdown-fetch-bubbles.scale, td_fetch_bubbles_scale_slm, "2");
+/* uops_retired.all */
+EVENT_ATTR_STR(topdown-slots-issued, td_slots_issued_slm,
+ "event=0xc2,umask=0x10");
+/* uops_retired.all */
+EVENT_ATTR_STR(topdown-slots-retired, td_slots_retired_slm,
+ "event=0xc2,umask=0x10");
+
+static struct attribute *slm_events_attrs[] = {
+ EVENT_PTR(td_total_slots_slm),
+ EVENT_PTR(td_total_slots_scale_slm),
+ EVENT_PTR(td_fetch_bubbles_slm),
+ EVENT_PTR(td_fetch_bubbles_scale_slm),
+ EVENT_PTR(td_slots_issued_slm),
+ EVENT_PTR(td_slots_retired_slm),
+ NULL
+};
+
static struct extra_reg intel_slm_extra_regs[] __read_mostly =
{
/* must define OFFCORE_RSP_X first, see intel_fixup_er() */
@@ -1671,9 +1730,11 @@ static __initconst const u64 knl_hw_cache_extra_regs
* disabled state if called consecutively.
*
* During consecutive calls, the same disable value will be written to related
- * registers, so the PMU state remains unchanged. hw.state in
- * intel_bts_disable_local will remain PERF_HES_STOPPED too in consecutive
- * calls.
+ * registers, so the PMU state remains unchanged.
+ *
+ * intel_bts events don't coexist with intel PMU's BTS events because of
+ * x86_add_exclusive(x86_lbr_exclusive_lbr); there's no need to keep them
+ * disabled around intel PMU's event batching etc, only inside the PMI handler.
*/
static void __intel_pmu_disable_all(void)
{
@@ -1683,8 +1744,6 @@ static void __intel_pmu_disable_all(void)
if (test_bit(INTEL_PMC_IDX_FIXED_BTS, cpuc->active_mask))
intel_pmu_disable_bts();
- else
- intel_bts_disable_local();
intel_pmu_pebs_disable_all();
}
@@ -1712,8 +1771,7 @@ static void __intel_pmu_enable_all(int added, bool pmi)
return;
intel_pmu_enable_bts(event->hw.config);
- } else
- intel_bts_enable_local();
+ }
}
static void intel_pmu_enable_all(int added)
@@ -2014,6 +2072,7 @@ static int intel_pmu_handle_irq(struct pt_regs *regs)
*/
if (!x86_pmu.late_ack)
apic_write(APIC_LVTPC, APIC_DM_NMI);
+ intel_bts_disable_local();
__intel_pmu_disable_all();
handled = intel_pmu_drain_bts_buffer();
handled += intel_bts_interrupt();
@@ -2113,6 +2172,7 @@ done:
/* Only restore PMU state when it's active. See x86_pmu_disable(). */
if (cpuc->enabled)
__intel_pmu_enable_all(0, true);
+ intel_bts_enable_local();
/*
* Only unmask the NMI after the overflow counters
@@ -3050,7 +3110,7 @@ static int intel_pmu_cpu_prepare(int cpu)
cpuc->excl_thread_id = 0;
}
- return NOTIFY_OK;
+ return 0;
err_constraint_list:
kfree(cpuc->constraint_list);
@@ -3061,7 +3121,7 @@ err_shared_regs:
cpuc->shared_regs = NULL;
err:
- return NOTIFY_BAD;
+ return -ENOMEM;
}
static void intel_pmu_cpu_starting(int cpu)
@@ -3290,11 +3350,11 @@ static int intel_snb_pebs_broken(int cpu)
u32 rev = UINT_MAX; /* default to broken for unknown models */
switch (cpu_data(cpu).x86_model) {
- case 42: /* SNB */
+ case INTEL_FAM6_SANDYBRIDGE:
rev = 0x28;
break;
- case 45: /* SNB-EP */
+ case INTEL_FAM6_SANDYBRIDGE_X:
switch (cpu_data(cpu).x86_mask) {
case 6: rev = 0x618; break;
case 7: rev = 0x70c; break;
@@ -3331,6 +3391,13 @@ static void intel_snb_check_microcode(void)
}
}
+static bool is_lbr_from(unsigned long msr)
+{
+ unsigned long lbr_from_nr = x86_pmu.lbr_from + x86_pmu.lbr_nr;
+
+ return x86_pmu.lbr_from <= msr && msr < lbr_from_nr;
+}
+
/*
* Under certain circumstances, access certain MSR may cause #GP.
* The function tests if the input MSR can be safely accessed.
@@ -3351,13 +3418,24 @@ static bool check_msr(unsigned long msr, u64 mask)
* Only change the bits which can be updated by wrmsrl.
*/
val_tmp = val_old ^ mask;
+
+ if (is_lbr_from(msr))
+ val_tmp = lbr_from_signext_quirk_wr(val_tmp);
+
if (wrmsrl_safe(msr, val_tmp) ||
rdmsrl_safe(msr, &val_new))
return false;
+ /*
+ * Quirk only affects validation in wrmsr(), so wrmsrl()'s value
+ * should equal rdmsrl()'s even with the quirk.
+ */
if (val_new != val_tmp)
return false;
+ if (is_lbr_from(msr))
+ val_old = lbr_from_signext_quirk_wr(val_old);
+
/* Here it's sure that the MSR can be safely accessed.
* Restore the old value and return.
*/
@@ -3466,6 +3544,13 @@ static struct attribute *hsw_events_attrs[] = {
EVENT_PTR(cycles_ct),
EVENT_PTR(mem_ld_hsw),
EVENT_PTR(mem_st_hsw),
+ EVENT_PTR(td_slots_issued),
+ EVENT_PTR(td_slots_retired),
+ EVENT_PTR(td_fetch_bubbles),
+ EVENT_PTR(td_total_slots),
+ EVENT_PTR(td_total_slots_scale),
+ EVENT_PTR(td_recovery_bubbles),
+ EVENT_PTR(td_recovery_bubbles_scale),
NULL
};
@@ -3537,15 +3622,15 @@ __init int intel_pmu_init(void)
* Install the hw-cache-events table:
*/
switch (boot_cpu_data.x86_model) {
- case 14: /* 65nm Core "Yonah" */
+ case INTEL_FAM6_CORE_YONAH:
pr_cont("Core events, ");
break;
- case 15: /* 65nm Core2 "Merom" */
+ case INTEL_FAM6_CORE2_MEROM:
x86_add_quirk(intel_clovertown_quirk);
- case 22: /* 65nm Core2 "Merom-L" */
- case 23: /* 45nm Core2 "Penryn" */
- case 29: /* 45nm Core2 "Dunnington (MP) */
+ case INTEL_FAM6_CORE2_MEROM_L:
+ case INTEL_FAM6_CORE2_PENRYN:
+ case INTEL_FAM6_CORE2_DUNNINGTON:
memcpy(hw_cache_event_ids, core2_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
@@ -3556,9 +3641,9 @@ __init int intel_pmu_init(void)
pr_cont("Core2 events, ");
break;
- case 30: /* 45nm Nehalem */
- case 26: /* 45nm Nehalem-EP */
- case 46: /* 45nm Nehalem-EX */
+ case INTEL_FAM6_NEHALEM:
+ case INTEL_FAM6_NEHALEM_EP:
+ case INTEL_FAM6_NEHALEM_EX:
memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
memcpy(hw_cache_extra_regs, nehalem_hw_cache_extra_regs,
@@ -3586,11 +3671,11 @@ __init int intel_pmu_init(void)
pr_cont("Nehalem events, ");
break;
- case 28: /* 45nm Atom "Pineview" */
- case 38: /* 45nm Atom "Lincroft" */
- case 39: /* 32nm Atom "Penwell" */
- case 53: /* 32nm Atom "Cloverview" */
- case 54: /* 32nm Atom "Cedarview" */
+ case INTEL_FAM6_ATOM_PINEVIEW:
+ case INTEL_FAM6_ATOM_LINCROFT:
+ case INTEL_FAM6_ATOM_PENWELL:
+ case INTEL_FAM6_ATOM_CLOVERVIEW:
+ case INTEL_FAM6_ATOM_CEDARVIEW:
memcpy(hw_cache_event_ids, atom_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
@@ -3602,9 +3687,9 @@ __init int intel_pmu_init(void)
pr_cont("Atom events, ");
break;
- case 55: /* 22nm Atom "Silvermont" */
- case 76: /* 14nm Atom "Airmont" */
- case 77: /* 22nm Atom "Silvermont Avoton/Rangely" */
+ case INTEL_FAM6_ATOM_SILVERMONT1:
+ case INTEL_FAM6_ATOM_SILVERMONT2:
+ case INTEL_FAM6_ATOM_AIRMONT:
memcpy(hw_cache_event_ids, slm_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
memcpy(hw_cache_extra_regs, slm_hw_cache_extra_regs,
@@ -3616,11 +3701,12 @@ __init int intel_pmu_init(void)
x86_pmu.pebs_constraints = intel_slm_pebs_event_constraints;
x86_pmu.extra_regs = intel_slm_extra_regs;
x86_pmu.flags |= PMU_FL_HAS_RSP_1;
+ x86_pmu.cpu_events = slm_events_attrs;
pr_cont("Silvermont events, ");
break;
- case 92: /* 14nm Atom "Goldmont" */
- case 95: /* 14nm Atom "Goldmont Denverton" */
+ case INTEL_FAM6_ATOM_GOLDMONT:
+ case INTEL_FAM6_ATOM_DENVERTON:
memcpy(hw_cache_event_ids, glm_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
memcpy(hw_cache_extra_regs, glm_hw_cache_extra_regs,
@@ -3643,9 +3729,9 @@ __init int intel_pmu_init(void)
pr_cont("Goldmont events, ");
break;
- case 37: /* 32nm Westmere */
- case 44: /* 32nm Westmere-EP */
- case 47: /* 32nm Westmere-EX */
+ case INTEL_FAM6_WESTMERE:
+ case INTEL_FAM6_WESTMERE_EP:
+ case INTEL_FAM6_WESTMERE_EX:
memcpy(hw_cache_event_ids, westmere_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
memcpy(hw_cache_extra_regs, nehalem_hw_cache_extra_regs,
@@ -3672,8 +3758,8 @@ __init int intel_pmu_init(void)
pr_cont("Westmere events, ");
break;
- case 42: /* 32nm SandyBridge */
- case 45: /* 32nm SandyBridge-E/EN/EP */
+ case INTEL_FAM6_SANDYBRIDGE:
+ case INTEL_FAM6_SANDYBRIDGE_X:
x86_add_quirk(intel_sandybridge_quirk);
x86_add_quirk(intel_ht_bug);
memcpy(hw_cache_event_ids, snb_hw_cache_event_ids,
@@ -3686,7 +3772,7 @@ __init int intel_pmu_init(void)
x86_pmu.event_constraints = intel_snb_event_constraints;
x86_pmu.pebs_constraints = intel_snb_pebs_event_constraints;
x86_pmu.pebs_aliases = intel_pebs_aliases_snb;
- if (boot_cpu_data.x86_model == 45)
+ if (boot_cpu_data.x86_model == INTEL_FAM6_SANDYBRIDGE_X)
x86_pmu.extra_regs = intel_snbep_extra_regs;
else
x86_pmu.extra_regs = intel_snb_extra_regs;
@@ -3708,8 +3794,8 @@ __init int intel_pmu_init(void)
pr_cont("SandyBridge events, ");
break;
- case 58: /* 22nm IvyBridge */
- case 62: /* 22nm IvyBridge-EP/EX */
+ case INTEL_FAM6_IVYBRIDGE:
+ case INTEL_FAM6_IVYBRIDGE_X:
x86_add_quirk(intel_ht_bug);
memcpy(hw_cache_event_ids, snb_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
@@ -3725,7 +3811,7 @@ __init int intel_pmu_init(void)
x86_pmu.pebs_constraints = intel_ivb_pebs_event_constraints;
x86_pmu.pebs_aliases = intel_pebs_aliases_ivb;
x86_pmu.pebs_prec_dist = true;
- if (boot_cpu_data.x86_model == 62)
+ if (boot_cpu_data.x86_model == INTEL_FAM6_IVYBRIDGE_X)
x86_pmu.extra_regs = intel_snbep_extra_regs;
else
x86_pmu.extra_regs = intel_snb_extra_regs;
@@ -3743,10 +3829,10 @@ __init int intel_pmu_init(void)
break;
- case 60: /* 22nm Haswell Core */
- case 63: /* 22nm Haswell Server */
- case 69: /* 22nm Haswell ULT */
- case 70: /* 22nm Haswell + GT3e (Intel Iris Pro graphics) */
+ case INTEL_FAM6_HASWELL_CORE:
+ case INTEL_FAM6_HASWELL_X:
+ case INTEL_FAM6_HASWELL_ULT:
+ case INTEL_FAM6_HASWELL_GT3E:
x86_add_quirk(intel_ht_bug);
x86_pmu.late_ack = true;
memcpy(hw_cache_event_ids, hsw_hw_cache_event_ids, sizeof(hw_cache_event_ids));
@@ -3770,10 +3856,10 @@ __init int intel_pmu_init(void)
pr_cont("Haswell events, ");
break;
- case 61: /* 14nm Broadwell Core-M */
- case 86: /* 14nm Broadwell Xeon D */
- case 71: /* 14nm Broadwell + GT3e (Intel Iris Pro graphics) */
- case 79: /* 14nm Broadwell Server */
+ case INTEL_FAM6_BROADWELL_CORE:
+ case INTEL_FAM6_BROADWELL_XEON_D:
+ case INTEL_FAM6_BROADWELL_GT3E:
+ case INTEL_FAM6_BROADWELL_X:
x86_pmu.late_ack = true;
memcpy(hw_cache_event_ids, hsw_hw_cache_event_ids, sizeof(hw_cache_event_ids));
memcpy(hw_cache_extra_regs, hsw_hw_cache_extra_regs, sizeof(hw_cache_extra_regs));
@@ -3806,7 +3892,7 @@ __init int intel_pmu_init(void)
pr_cont("Broadwell events, ");
break;
- case 87: /* Knights Landing Xeon Phi */
+ case INTEL_FAM6_XEON_PHI_KNL:
memcpy(hw_cache_event_ids,
slm_hw_cache_event_ids, sizeof(hw_cache_event_ids));
memcpy(hw_cache_extra_regs,
@@ -3824,16 +3910,22 @@ __init int intel_pmu_init(void)
pr_cont("Knights Landing events, ");
break;
- case 142: /* 14nm Kabylake Mobile */
- case 158: /* 14nm Kabylake Desktop */
- case 78: /* 14nm Skylake Mobile */
- case 94: /* 14nm Skylake Desktop */
- case 85: /* 14nm Skylake Server */
+ case INTEL_FAM6_SKYLAKE_MOBILE:
+ case INTEL_FAM6_SKYLAKE_DESKTOP:
+ case INTEL_FAM6_SKYLAKE_X:
+ case INTEL_FAM6_KABYLAKE_MOBILE:
+ case INTEL_FAM6_KABYLAKE_DESKTOP:
x86_pmu.late_ack = true;
memcpy(hw_cache_event_ids, skl_hw_cache_event_ids, sizeof(hw_cache_event_ids));
memcpy(hw_cache_extra_regs, skl_hw_cache_extra_regs, sizeof(hw_cache_extra_regs));
intel_pmu_lbr_init_skl();
+ /* INT_MISC.RECOVERY_CYCLES has umask 1 in Skylake */
+ event_attr_td_recovery_bubbles.event_str_noht =
+ "event=0xd,umask=0x1,cmask=1";
+ event_attr_td_recovery_bubbles.event_str_ht =
+ "event=0xd,umask=0x1,cmask=1,any=1";
+
x86_pmu.event_constraints = intel_skl_event_constraints;
x86_pmu.pebs_constraints = intel_skl_pebs_event_constraints;
x86_pmu.extra_regs = intel_skl_extra_regs;
@@ -3914,6 +4006,8 @@ __init int intel_pmu_init(void)
x86_pmu.lbr_nr = 0;
}
+ if (x86_pmu.lbr_nr)
+ pr_cont("%d-deep LBR, ", x86_pmu.lbr_nr);
/*
* Access extra MSR may cause #GP under certain circumstances.
* E.g. KVM doesn't support offcore event
@@ -3946,16 +4040,14 @@ __init int intel_pmu_init(void)
*/
static __init int fixup_ht_bug(void)
{
- int cpu = smp_processor_id();
- int w, c;
+ int c;
/*
* problem not present on this CPU model, nothing to do
*/
if (!(x86_pmu.flags & PMU_FL_EXCL_ENABLED))
return 0;
- w = cpumask_weight(topology_sibling_cpumask(cpu));
- if (w > 1) {
+ if (topology_max_smt_threads() > 1) {
pr_info("PMU erratum BJ122, BV98, HSD29 worked around, HT is on\n");
return 0;
}
diff --git a/arch/x86/events/intel/cqm.c b/arch/x86/events/intel/cqm.c
index 4ff41ae51..8f82b0293 100644
--- a/arch/x86/events/intel/cqm.c
+++ b/arch/x86/events/intel/cqm.c
@@ -1586,7 +1586,7 @@ static inline void cqm_pick_event_reader(int cpu)
cpumask_set_cpu(cpu, &cqm_cpumask);
}
-static void intel_cqm_cpu_starting(unsigned int cpu)
+static int intel_cqm_cpu_starting(unsigned int cpu)
{
struct intel_pqr_state *state = &per_cpu(pqr_state, cpu);
struct cpuinfo_x86 *c = &cpu_data(cpu);
@@ -1597,39 +1597,26 @@ static void intel_cqm_cpu_starting(unsigned int cpu)
WARN_ON(c->x86_cache_max_rmid != cqm_max_rmid);
WARN_ON(c->x86_cache_occ_scale != cqm_l3_scale);
+
+ cqm_pick_event_reader(cpu);
+ return 0;
}
-static void intel_cqm_cpu_exit(unsigned int cpu)
+static int intel_cqm_cpu_exit(unsigned int cpu)
{
int target;
/* Is @cpu the current cqm reader for this package ? */
if (!cpumask_test_and_clear_cpu(cpu, &cqm_cpumask))
- return;
+ return 0;
/* Find another online reader in this package */
target = cpumask_any_but(topology_core_cpumask(cpu), cpu);
if (target < nr_cpu_ids)
cpumask_set_cpu(target, &cqm_cpumask);
-}
-
-static int intel_cqm_cpu_notifier(struct notifier_block *nb,
- unsigned long action, void *hcpu)
-{
- unsigned int cpu = (unsigned long)hcpu;
-
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_DOWN_PREPARE:
- intel_cqm_cpu_exit(cpu);
- break;
- case CPU_STARTING:
- intel_cqm_cpu_starting(cpu);
- cqm_pick_event_reader(cpu);
- break;
- }
- return NOTIFY_OK;
+ return 0;
}
static const struct x86_cpu_id intel_cqm_match[] = {
@@ -1691,7 +1678,7 @@ out:
static int __init intel_cqm_init(void)
{
char *str = NULL, scale[20];
- int i, cpu, ret;
+ int cpu, ret;
if (x86_match_cpu(intel_cqm_match))
cqm_enabled = true;
@@ -1714,8 +1701,7 @@ static int __init intel_cqm_init(void)
*
* Also, check that the scales match on all cpus.
*/
- cpu_notifier_register_begin();
-
+ get_online_cpus();
for_each_online_cpu(cpu) {
struct cpuinfo_x86 *c = &cpu_data(cpu);
@@ -1752,11 +1738,6 @@ static int __init intel_cqm_init(void)
if (ret)
goto out;
- for_each_online_cpu(i) {
- intel_cqm_cpu_starting(i);
- cqm_pick_event_reader(i);
- }
-
if (mbm_enabled)
ret = intel_mbm_init();
if (ret && !cqm_enabled)
@@ -1781,12 +1762,18 @@ static int __init intel_cqm_init(void)
pr_info("Intel MBM enabled\n");
/*
- * Register the hot cpu notifier once we are sure cqm
+ * Setup the hot cpu notifier once we are sure cqm
* is enabled to avoid notifier leak.
*/
- __perf_cpu_notifier(intel_cqm_cpu_notifier);
+ cpuhp_setup_state(CPUHP_AP_PERF_X86_CQM_STARTING,
+ "AP_PERF_X86_CQM_STARTING",
+ intel_cqm_cpu_starting, NULL);
+ cpuhp_setup_state(CPUHP_AP_PERF_X86_CQM_ONLINE, "AP_PERF_X86_CQM_ONLINE",
+ NULL, intel_cqm_cpu_exit);
+
out:
- cpu_notifier_register_done();
+ put_online_cpus();
+
if (ret) {
kfree(str);
cqm_cleanup();
diff --git a/arch/x86/events/intel/cstate.c b/arch/x86/events/intel/cstate.c
index 9ba4e4136..3ca87b5a8 100644
--- a/arch/x86/events/intel/cstate.c
+++ b/arch/x86/events/intel/cstate.c
@@ -89,6 +89,7 @@
#include <linux/slab.h>
#include <linux/perf_event.h>
#include <asm/cpu_device_id.h>
+#include <asm/intel-family.h>
#include "../perf_event.h"
MODULE_LICENSE("GPL");
@@ -365,7 +366,7 @@ static int cstate_pmu_event_add(struct perf_event *event, int mode)
* Check if exiting cpu is the designated reader. If so migrate the
* events when there is a valid target available
*/
-static void cstate_cpu_exit(int cpu)
+static int cstate_cpu_exit(unsigned int cpu)
{
unsigned int target;
@@ -390,9 +391,10 @@ static void cstate_cpu_exit(int cpu)
perf_pmu_migrate_context(&cstate_pkg_pmu, cpu, target);
}
}
+ return 0;
}
-static void cstate_cpu_init(int cpu)
+static int cstate_cpu_init(unsigned int cpu)
{
unsigned int target;
@@ -414,31 +416,10 @@ static void cstate_cpu_init(int cpu)
topology_core_cpumask(cpu));
if (has_cstate_pkg && target >= nr_cpu_ids)
cpumask_set_cpu(cpu, &cstate_pkg_cpu_mask);
-}
-static int cstate_cpu_notifier(struct notifier_block *self,
- unsigned long action, void *hcpu)
-{
- unsigned int cpu = (long)hcpu;
-
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_STARTING:
- cstate_cpu_init(cpu);
- break;
- case CPU_DOWN_PREPARE:
- cstate_cpu_exit(cpu);
- break;
- default:
- break;
- }
- return NOTIFY_OK;
+ return 0;
}
-static struct notifier_block cstate_cpu_nb = {
- .notifier_call = cstate_cpu_notifier,
- .priority = CPU_PRI_PERF + 1,
-};
-
static struct pmu cstate_core_pmu = {
.attr_groups = core_attr_groups,
.name = "cstate_core",
@@ -511,37 +492,37 @@ static const struct cstate_model slm_cstates __initconst = {
{ X86_VENDOR_INTEL, 6, model, X86_FEATURE_ANY, (unsigned long) &(states) }
static const struct x86_cpu_id intel_cstates_match[] __initconst = {
- X86_CSTATES_MODEL(30, nhm_cstates), /* 45nm Nehalem */
- X86_CSTATES_MODEL(26, nhm_cstates), /* 45nm Nehalem-EP */
- X86_CSTATES_MODEL(46, nhm_cstates), /* 45nm Nehalem-EX */
+ X86_CSTATES_MODEL(INTEL_FAM6_NEHALEM, nhm_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_NEHALEM_EP, nhm_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_NEHALEM_EX, nhm_cstates),
- X86_CSTATES_MODEL(37, nhm_cstates), /* 32nm Westmere */
- X86_CSTATES_MODEL(44, nhm_cstates), /* 32nm Westmere-EP */
- X86_CSTATES_MODEL(47, nhm_cstates), /* 32nm Westmere-EX */
+ X86_CSTATES_MODEL(INTEL_FAM6_WESTMERE, nhm_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_WESTMERE_EP, nhm_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_WESTMERE_EX, nhm_cstates),
- X86_CSTATES_MODEL(42, snb_cstates), /* 32nm SandyBridge */
- X86_CSTATES_MODEL(45, snb_cstates), /* 32nm SandyBridge-E/EN/EP */
+ X86_CSTATES_MODEL(INTEL_FAM6_SANDYBRIDGE, snb_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_SANDYBRIDGE_X, snb_cstates),
- X86_CSTATES_MODEL(58, snb_cstates), /* 22nm IvyBridge */
- X86_CSTATES_MODEL(62, snb_cstates), /* 22nm IvyBridge-EP/EX */
+ X86_CSTATES_MODEL(INTEL_FAM6_IVYBRIDGE, snb_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_IVYBRIDGE_X, snb_cstates),
- X86_CSTATES_MODEL(60, snb_cstates), /* 22nm Haswell Core */
- X86_CSTATES_MODEL(63, snb_cstates), /* 22nm Haswell Server */
- X86_CSTATES_MODEL(70, snb_cstates), /* 22nm Haswell + GT3e */
+ X86_CSTATES_MODEL(INTEL_FAM6_HASWELL_CORE, snb_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_HASWELL_X, snb_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_HASWELL_GT3E, snb_cstates),
- X86_CSTATES_MODEL(69, hswult_cstates), /* 22nm Haswell ULT */
+ X86_CSTATES_MODEL(INTEL_FAM6_HASWELL_ULT, hswult_cstates),
- X86_CSTATES_MODEL(55, slm_cstates), /* 22nm Atom Silvermont */
- X86_CSTATES_MODEL(77, slm_cstates), /* 22nm Atom Avoton/Rangely */
- X86_CSTATES_MODEL(76, slm_cstates), /* 22nm Atom Airmont */
+ X86_CSTATES_MODEL(INTEL_FAM6_ATOM_SILVERMONT1, slm_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_ATOM_SILVERMONT2, slm_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_ATOM_AIRMONT, slm_cstates),
- X86_CSTATES_MODEL(61, snb_cstates), /* 14nm Broadwell Core-M */
- X86_CSTATES_MODEL(86, snb_cstates), /* 14nm Broadwell Xeon D */
- X86_CSTATES_MODEL(71, snb_cstates), /* 14nm Broadwell + GT3e */
- X86_CSTATES_MODEL(79, snb_cstates), /* 14nm Broadwell Server */
+ X86_CSTATES_MODEL(INTEL_FAM6_BROADWELL_CORE, snb_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_BROADWELL_XEON_D, snb_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_BROADWELL_GT3E, snb_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_BROADWELL_X, snb_cstates),
- X86_CSTATES_MODEL(78, snb_cstates), /* 14nm Skylake Mobile */
- X86_CSTATES_MODEL(94, snb_cstates), /* 14nm Skylake Desktop */
+ X86_CSTATES_MODEL(INTEL_FAM6_SKYLAKE_MOBILE, snb_cstates),
+ X86_CSTATES_MODEL(INTEL_FAM6_SKYLAKE_DESKTOP, snb_cstates),
{ },
};
MODULE_DEVICE_TABLE(x86cpu, intel_cstates_match);
@@ -599,18 +580,20 @@ static inline void cstate_cleanup(void)
static int __init cstate_init(void)
{
- int cpu, err;
+ int err;
- cpu_notifier_register_begin();
- for_each_online_cpu(cpu)
- cstate_cpu_init(cpu);
+ cpuhp_setup_state(CPUHP_AP_PERF_X86_CSTATE_STARTING,
+ "AP_PERF_X86_CSTATE_STARTING", cstate_cpu_init,
+ NULL);
+ cpuhp_setup_state(CPUHP_AP_PERF_X86_CSTATE_ONLINE,
+ "AP_PERF_X86_CSTATE_ONLINE", NULL, cstate_cpu_exit);
if (has_cstate_core) {
err = perf_pmu_register(&cstate_core_pmu, cstate_core_pmu.name, -1);
if (err) {
has_cstate_core = false;
pr_info("Failed to register cstate core pmu\n");
- goto out;
+ return err;
}
}
@@ -620,12 +603,10 @@ static int __init cstate_init(void)
has_cstate_pkg = false;
pr_info("Failed to register cstate pkg pmu\n");
cstate_cleanup();
- goto out;
+ return err;
}
}
- __register_cpu_notifier(&cstate_cpu_nb);
-out:
- cpu_notifier_register_done();
+
return err;
}
@@ -651,9 +632,8 @@ module_init(cstate_pmu_init);
static void __exit cstate_pmu_exit(void)
{
- cpu_notifier_register_begin();
- __unregister_cpu_notifier(&cstate_cpu_nb);
+ cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_CSTATE_ONLINE);
+ cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_CSTATE_STARTING);
cstate_cleanup();
- cpu_notifier_register_done();
}
module_exit(cstate_pmu_exit);
diff --git a/arch/x86/events/intel/lbr.c b/arch/x86/events/intel/lbr.c
index 9e2b40cdb..707d358e0 100644
--- a/arch/x86/events/intel/lbr.c
+++ b/arch/x86/events/intel/lbr.c
@@ -77,9 +77,11 @@ static enum {
LBR_IND_JMP |\
LBR_FAR)
-#define LBR_FROM_FLAG_MISPRED (1ULL << 63)
-#define LBR_FROM_FLAG_IN_TX (1ULL << 62)
-#define LBR_FROM_FLAG_ABORT (1ULL << 61)
+#define LBR_FROM_FLAG_MISPRED BIT_ULL(63)
+#define LBR_FROM_FLAG_IN_TX BIT_ULL(62)
+#define LBR_FROM_FLAG_ABORT BIT_ULL(61)
+
+#define LBR_FROM_SIGNEXT_2MSB (BIT_ULL(60) | BIT_ULL(59))
/*
* x86control flow change classification
@@ -235,6 +237,97 @@ enum {
LBR_VALID,
};
+/*
+ * For formats with LBR_TSX flags (e.g. LBR_FORMAT_EIP_FLAGS2), bits 61:62 in
+ * MSR_LAST_BRANCH_FROM_x are the TSX flags when TSX is supported, but when
+ * TSX is not supported they have no consistent behavior:
+ *
+ * - For wrmsr(), bits 61:62 are considered part of the sign extension.
+ * - For HW updates (branch captures) bits 61:62 are always OFF and are not
+ * part of the sign extension.
+ *
+ * Therefore, if:
+ *
+ * 1) LBR has TSX format
+ * 2) CPU has no TSX support enabled
+ *
+ * ... then any value passed to wrmsr() must be sign extended to 63 bits and any
+ * value from rdmsr() must be converted to have a 61 bits sign extension,
+ * ignoring the TSX flags.
+ */
+static inline bool lbr_from_signext_quirk_needed(void)
+{
+ int lbr_format = x86_pmu.intel_cap.lbr_format;
+ bool tsx_support = boot_cpu_has(X86_FEATURE_HLE) ||
+ boot_cpu_has(X86_FEATURE_RTM);
+
+ return !tsx_support && (lbr_desc[lbr_format] & LBR_TSX);
+}
+
+DEFINE_STATIC_KEY_FALSE(lbr_from_quirk_key);
+
+/* If quirk is enabled, ensure sign extension is 63 bits: */
+inline u64 lbr_from_signext_quirk_wr(u64 val)
+{
+ if (static_branch_unlikely(&lbr_from_quirk_key)) {
+ /*
+ * Sign extend into bits 61:62 while preserving bit 63.
+ *
+ * Quirk is enabled when TSX is disabled. Therefore TSX bits
+ * in val are always OFF and must be changed to be sign
+ * extension bits. Since bits 59:60 are guaranteed to be
+ * part of the sign extension bits, we can just copy them
+ * to 61:62.
+ */
+ val |= (LBR_FROM_SIGNEXT_2MSB & val) << 2;
+ }
+ return val;
+}
+
+/*
+ * If quirk is needed, ensure sign extension is 61 bits:
+ */
+u64 lbr_from_signext_quirk_rd(u64 val)
+{
+ if (static_branch_unlikely(&lbr_from_quirk_key)) {
+ /*
+ * Quirk is on when TSX is not enabled. Therefore TSX
+ * flags must be read as OFF.
+ */
+ val &= ~(LBR_FROM_FLAG_IN_TX | LBR_FROM_FLAG_ABORT);
+ }
+ return val;
+}
+
+static inline void wrlbr_from(unsigned int idx, u64 val)
+{
+ val = lbr_from_signext_quirk_wr(val);
+ wrmsrl(x86_pmu.lbr_from + idx, val);
+}
+
+static inline void wrlbr_to(unsigned int idx, u64 val)
+{
+ wrmsrl(x86_pmu.lbr_to + idx, val);
+}
+
+static inline u64 rdlbr_from(unsigned int idx)
+{
+ u64 val;
+
+ rdmsrl(x86_pmu.lbr_from + idx, val);
+
+ return lbr_from_signext_quirk_rd(val);
+}
+
+static inline u64 rdlbr_to(unsigned int idx)
+{
+ u64 val;
+
+ rdmsrl(x86_pmu.lbr_to + idx, val);
+
+ return val;
+}
+
static void __intel_pmu_lbr_restore(struct x86_perf_task_context *task_ctx)
{
int i;
@@ -251,8 +344,9 @@ static void __intel_pmu_lbr_restore(struct x86_perf_task_context *task_ctx)
tos = task_ctx->tos;
for (i = 0; i < tos; i++) {
lbr_idx = (tos - i) & mask;
- wrmsrl(x86_pmu.lbr_from + lbr_idx, task_ctx->lbr_from[i]);
- wrmsrl(x86_pmu.lbr_to + lbr_idx, task_ctx->lbr_to[i]);
+ wrlbr_from(lbr_idx, task_ctx->lbr_from[i]);
+ wrlbr_to (lbr_idx, task_ctx->lbr_to[i]);
+
if (x86_pmu.intel_cap.lbr_format == LBR_FORMAT_INFO)
wrmsrl(MSR_LBR_INFO_0 + lbr_idx, task_ctx->lbr_info[i]);
}
@@ -262,9 +356,9 @@ static void __intel_pmu_lbr_restore(struct x86_perf_task_context *task_ctx)
static void __intel_pmu_lbr_save(struct x86_perf_task_context *task_ctx)
{
- int i;
unsigned lbr_idx, mask;
u64 tos;
+ int i;
if (task_ctx->lbr_callstack_users == 0) {
task_ctx->lbr_stack_state = LBR_NONE;
@@ -275,8 +369,8 @@ static void __intel_pmu_lbr_save(struct x86_perf_task_context *task_ctx)
tos = intel_pmu_lbr_tos();
for (i = 0; i < tos; i++) {
lbr_idx = (tos - i) & mask;
- rdmsrl(x86_pmu.lbr_from + lbr_idx, task_ctx->lbr_from[i]);
- rdmsrl(x86_pmu.lbr_to + lbr_idx, task_ctx->lbr_to[i]);
+ task_ctx->lbr_from[i] = rdlbr_from(lbr_idx);
+ task_ctx->lbr_to[i] = rdlbr_to(lbr_idx);
if (x86_pmu.intel_cap.lbr_format == LBR_FORMAT_INFO)
rdmsrl(MSR_LBR_INFO_0 + lbr_idx, task_ctx->lbr_info[i]);
}
@@ -452,8 +546,8 @@ static void intel_pmu_lbr_read_64(struct cpu_hw_events *cpuc)
u16 cycles = 0;
int lbr_flags = lbr_desc[lbr_format];
- rdmsrl(x86_pmu.lbr_from + lbr_idx, from);
- rdmsrl(x86_pmu.lbr_to + lbr_idx, to);
+ from = rdlbr_from(lbr_idx);
+ to = rdlbr_to(lbr_idx);
if (lbr_format == LBR_FORMAT_INFO && need_info) {
u64 info;
@@ -956,7 +1050,6 @@ void __init intel_pmu_lbr_init_core(void)
* SW branch filter usage:
* - compensate for lack of HW filter
*/
- pr_cont("4-deep LBR, ");
}
/* nehalem/westmere */
@@ -977,7 +1070,6 @@ void __init intel_pmu_lbr_init_nhm(void)
* That requires LBR_FAR but that means far
* jmp need to be filtered out
*/
- pr_cont("16-deep LBR, ");
}
/* sandy bridge */
@@ -997,7 +1089,6 @@ void __init intel_pmu_lbr_init_snb(void)
* That requires LBR_FAR but that means far
* jmp need to be filtered out
*/
- pr_cont("16-deep LBR, ");
}
/* haswell */
@@ -1011,7 +1102,8 @@ void intel_pmu_lbr_init_hsw(void)
x86_pmu.lbr_sel_mask = LBR_SEL_MASK;
x86_pmu.lbr_sel_map = hsw_lbr_sel_map;
- pr_cont("16-deep LBR, ");
+ if (lbr_from_signext_quirk_needed())
+ static_branch_enable(&lbr_from_quirk_key);
}
/* skylake */
@@ -1031,7 +1123,6 @@ __init void intel_pmu_lbr_init_skl(void)
* That requires LBR_FAR but that means far
* jmp need to be filtered out
*/
- pr_cont("32-deep LBR, ");
}
/* atom */
@@ -1057,7 +1148,6 @@ void __init intel_pmu_lbr_init_atom(void)
* SW branch filter usage:
* - compensate for lack of HW filter
*/
- pr_cont("8-deep LBR, ");
}
/* slm */
@@ -1088,6 +1178,4 @@ void intel_pmu_lbr_init_knl(void)
x86_pmu.lbr_sel_mask = LBR_SEL_MASK;
x86_pmu.lbr_sel_map = snb_lbr_sel_map;
-
- pr_cont("8-deep LBR, ");
}
diff --git a/arch/x86/events/intel/rapl.c b/arch/x86/events/intel/rapl.c
index e30eef4f2..28865938a 100644
--- a/arch/x86/events/intel/rapl.c
+++ b/arch/x86/events/intel/rapl.c
@@ -55,6 +55,7 @@
#include <linux/slab.h>
#include <linux/perf_event.h>
#include <asm/cpu_device_id.h>
+#include <asm/intel-family.h>
#include "../perf_event.h"
MODULE_LICENSE("GPL");
@@ -555,14 +556,14 @@ const struct attribute_group *rapl_attr_groups[] = {
NULL,
};
-static void rapl_cpu_exit(int cpu)
+static int rapl_cpu_offline(unsigned int cpu)
{
struct rapl_pmu *pmu = cpu_to_rapl_pmu(cpu);
int target;
/* Check if exiting cpu is used for collecting rapl events */
if (!cpumask_test_and_clear_cpu(cpu, &rapl_cpu_mask))
- return;
+ return 0;
pmu->cpu = -1;
/* Find a new cpu to collect rapl events */
@@ -574,9 +575,10 @@ static void rapl_cpu_exit(int cpu)
pmu->cpu = target;
perf_pmu_migrate_context(pmu->pmu, cpu, target);
}
+ return 0;
}
-static void rapl_cpu_init(int cpu)
+static int rapl_cpu_online(unsigned int cpu)
{
struct rapl_pmu *pmu = cpu_to_rapl_pmu(cpu);
int target;
@@ -587,13 +589,14 @@ static void rapl_cpu_init(int cpu)
*/
target = cpumask_any_and(&rapl_cpu_mask, topology_core_cpumask(cpu));
if (target < nr_cpu_ids)
- return;
+ return 0;
cpumask_set_cpu(cpu, &rapl_cpu_mask);
pmu->cpu = cpu;
+ return 0;
}
-static int rapl_cpu_prepare(int cpu)
+static int rapl_cpu_prepare(unsigned int cpu)
{
struct rapl_pmu *pmu = cpu_to_rapl_pmu(cpu);
@@ -614,33 +617,6 @@ static int rapl_cpu_prepare(int cpu)
return 0;
}
-static int rapl_cpu_notifier(struct notifier_block *self,
- unsigned long action, void *hcpu)
-{
- unsigned int cpu = (long)hcpu;
-
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_UP_PREPARE:
- rapl_cpu_prepare(cpu);
- break;
-
- case CPU_DOWN_FAILED:
- case CPU_ONLINE:
- rapl_cpu_init(cpu);
- break;
-
- case CPU_DOWN_PREPARE:
- rapl_cpu_exit(cpu);
- break;
- }
- return NOTIFY_OK;
-}
-
-static struct notifier_block rapl_cpu_nb = {
- .notifier_call = rapl_cpu_notifier,
- .priority = CPU_PRI_PERF + 1,
-};
-
static int rapl_check_hw_unit(bool apply_quirk)
{
u64 msr_rapl_power_unit_bits;
@@ -691,24 +667,6 @@ static void __init rapl_advertise(void)
}
}
-static int __init rapl_prepare_cpus(void)
-{
- unsigned int cpu, pkg;
- int ret;
-
- for_each_online_cpu(cpu) {
- pkg = topology_logical_package_id(cpu);
- if (rapl_pmus->pmus[pkg])
- continue;
-
- ret = rapl_cpu_prepare(cpu);
- if (ret)
- return ret;
- rapl_cpu_init(cpu);
- }
- return 0;
-}
-
static void cleanup_rapl_pmus(void)
{
int i;
@@ -786,26 +744,27 @@ static const struct intel_rapl_init_fun skl_rapl_init __initconst = {
};
static const struct x86_cpu_id rapl_cpu_match[] __initconst = {
- X86_RAPL_MODEL_MATCH(42, snb_rapl_init), /* Sandy Bridge */
- X86_RAPL_MODEL_MATCH(45, snbep_rapl_init), /* Sandy Bridge-EP */
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_SANDYBRIDGE, snb_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_SANDYBRIDGE_X, snbep_rapl_init),
- X86_RAPL_MODEL_MATCH(58, snb_rapl_init), /* Ivy Bridge */
- X86_RAPL_MODEL_MATCH(62, snbep_rapl_init), /* IvyTown */
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_IVYBRIDGE, snb_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_IVYBRIDGE_X, snbep_rapl_init),
- X86_RAPL_MODEL_MATCH(60, hsw_rapl_init), /* Haswell */
- X86_RAPL_MODEL_MATCH(63, hsx_rapl_init), /* Haswell-Server */
- X86_RAPL_MODEL_MATCH(69, hsw_rapl_init), /* Haswell-Celeron */
- X86_RAPL_MODEL_MATCH(70, hsw_rapl_init), /* Haswell GT3e */
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_HASWELL_CORE, hsw_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_HASWELL_X, hsw_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_HASWELL_ULT, hsw_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_HASWELL_GT3E, hsw_rapl_init),
- X86_RAPL_MODEL_MATCH(61, hsw_rapl_init), /* Broadwell */
- X86_RAPL_MODEL_MATCH(71, hsw_rapl_init), /* Broadwell-H */
- X86_RAPL_MODEL_MATCH(79, hsx_rapl_init), /* Broadwell-Server */
- X86_RAPL_MODEL_MATCH(86, hsx_rapl_init), /* Broadwell Xeon D */
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_BROADWELL_CORE, hsw_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_BROADWELL_GT3E, hsw_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_BROADWELL_X, hsw_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_BROADWELL_XEON_D, hsw_rapl_init),
- X86_RAPL_MODEL_MATCH(87, knl_rapl_init), /* Knights Landing */
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_XEON_PHI_KNL, knl_rapl_init),
- X86_RAPL_MODEL_MATCH(78, skl_rapl_init), /* Skylake */
- X86_RAPL_MODEL_MATCH(94, skl_rapl_init), /* Skylake H/S */
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_SKYLAKE_MOBILE, skl_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_SKYLAKE_DESKTOP, skl_rapl_init),
+ X86_RAPL_MODEL_MATCH(INTEL_FAM6_SKYLAKE_X, hsx_rapl_init),
{},
};
@@ -835,35 +794,44 @@ static int __init rapl_pmu_init(void)
if (ret)
return ret;
- cpu_notifier_register_begin();
+ /*
+ * Install callbacks. Core will call them for each online cpu.
+ */
- ret = rapl_prepare_cpus();
+ ret = cpuhp_setup_state(CPUHP_PERF_X86_RAPL_PREP, "PERF_X86_RAPL_PREP",
+ rapl_cpu_prepare, NULL);
if (ret)
goto out;
+ ret = cpuhp_setup_state(CPUHP_AP_PERF_X86_RAPL_ONLINE,
+ "AP_PERF_X86_RAPL_ONLINE",
+ rapl_cpu_online, rapl_cpu_offline);
+ if (ret)
+ goto out1;
+
ret = perf_pmu_register(&rapl_pmus->pmu, "power", -1);
if (ret)
- goto out;
+ goto out2;
- __register_cpu_notifier(&rapl_cpu_nb);
- cpu_notifier_register_done();
rapl_advertise();
return 0;
+out2:
+ cpuhp_remove_state(CPUHP_AP_PERF_X86_RAPL_ONLINE);
+out1:
+ cpuhp_remove_state(CPUHP_PERF_X86_RAPL_PREP);
out:
pr_warn("Initialization failed (%d), disabled\n", ret);
cleanup_rapl_pmus();
- cpu_notifier_register_done();
return ret;
}
module_init(rapl_pmu_init);
static void __exit intel_rapl_exit(void)
{
- cpu_notifier_register_begin();
- __unregister_cpu_notifier(&rapl_cpu_nb);
+ cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_RAPL_ONLINE);
+ cpuhp_remove_state_nocalls(CPUHP_PERF_X86_RAPL_PREP);
perf_pmu_unregister(&rapl_pmus->pmu);
cleanup_rapl_pmus();
- cpu_notifier_register_done();
}
module_exit(intel_rapl_exit);
diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index fce74062d..463dc7a5a 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -1,4 +1,7 @@
+#include <linux/module.h>
+
#include <asm/cpu_device_id.h>
+#include <asm/intel-family.h>
#include "uncore.h"
static struct intel_uncore_type *empty_uncore[] = { NULL, };
@@ -882,7 +885,7 @@ uncore_types_init(struct intel_uncore_type **types, bool setid)
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;
+ struct intel_uncore_pmu *pmu = NULL;
struct intel_uncore_box *box;
int phys_id, pkg, ret;
@@ -903,20 +906,37 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id
}
type = uncore_pci_uncores[UNCORE_PCI_DEV_TYPE(id->driver_data)];
+
/*
- * for performance monitoring unit with multiple boxes,
- * each box has a different function id.
- */
- pmu = &type->pmus[UNCORE_PCI_DEV_IDX(id->driver_data)];
- /* Knights Landing uses a common PCI device ID for multiple instances of
- * an uncore PMU device type. There is only one entry per device type in
- * the knl_uncore_pci_ids table inspite of multiple devices present for
- * some device types. Hence PCI device idx would be 0 for all devices.
- * So increment pmu pointer to point to an unused array element.
+ * Some platforms, e.g. Knights Landing, use a common PCI device ID
+ * for multiple instances of an uncore PMU device type. We should check
+ * PCI slot and func to indicate the uncore box.
*/
- if (boot_cpu_data.x86_model == 87) {
- while (pmu->func_id >= 0)
- pmu++;
+ if (id->driver_data & ~0xffff) {
+ struct pci_driver *pci_drv = pdev->driver;
+ const struct pci_device_id *ids = pci_drv->id_table;
+ unsigned int devfn;
+
+ while (ids && ids->vendor) {
+ if ((ids->vendor == pdev->vendor) &&
+ (ids->device == pdev->device)) {
+ devfn = PCI_DEVFN(UNCORE_PCI_DEV_DEV(ids->driver_data),
+ UNCORE_PCI_DEV_FUNC(ids->driver_data));
+ if (devfn == pdev->devfn) {
+ pmu = &type->pmus[UNCORE_PCI_DEV_IDX(ids->driver_data)];
+ break;
+ }
+ }
+ ids++;
+ }
+ if (pmu == NULL)
+ return -ENODEV;
+ } else {
+ /*
+ * for performance monitoring unit with multiple boxes,
+ * each box has a different function id.
+ */
+ pmu = &type->pmus[UNCORE_PCI_DEV_IDX(id->driver_data)];
}
if (WARN_ON_ONCE(pmu->boxes[pkg] != NULL))
@@ -956,7 +976,7 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id
static void uncore_pci_remove(struct pci_dev *pdev)
{
- struct intel_uncore_box *box = pci_get_drvdata(pdev);
+ struct intel_uncore_box *box;
struct intel_uncore_pmu *pmu;
int i, phys_id, pkg;
@@ -1034,7 +1054,7 @@ static void uncore_pci_exit(void)
}
}
-static void uncore_cpu_dying(int cpu)
+static int uncore_cpu_dying(unsigned int cpu)
{
struct intel_uncore_type *type, **types = uncore_msr_uncores;
struct intel_uncore_pmu *pmu;
@@ -1051,16 +1071,19 @@ static void uncore_cpu_dying(int cpu)
uncore_box_exit(box);
}
}
+ return 0;
}
-static void uncore_cpu_starting(int cpu, bool init)
+static int first_init;
+
+static int uncore_cpu_starting(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, ncpus = 1;
- if (init) {
+ if (first_init) {
/*
* On init we get the number of online cpus in the package
* and set refcount for all of them.
@@ -1081,9 +1104,11 @@ static void uncore_cpu_starting(int cpu, bool init)
uncore_box_init(box);
}
}
+
+ return 0;
}
-static int uncore_cpu_prepare(int cpu)
+static int uncore_cpu_prepare(unsigned int cpu)
{
struct intel_uncore_type *type, **types = uncore_msr_uncores;
struct intel_uncore_pmu *pmu;
@@ -1146,13 +1171,13 @@ static void uncore_change_context(struct intel_uncore_type **uncores,
uncore_change_type_ctx(*uncores, old_cpu, new_cpu);
}
-static void uncore_event_exit_cpu(int cpu)
+static int uncore_event_cpu_offline(unsigned int cpu)
{
int target;
/* Check if exiting cpu is used for collecting uncore events */
if (!cpumask_test_and_clear_cpu(cpu, &uncore_cpu_mask))
- return;
+ return 0;
/* Find a new cpu to collect uncore events */
target = cpumask_any_but(topology_core_cpumask(cpu), cpu);
@@ -1165,9 +1190,10 @@ static void uncore_event_exit_cpu(int cpu)
uncore_change_context(uncore_msr_uncores, cpu, target);
uncore_change_context(uncore_pci_uncores, cpu, target);
+ return 0;
}
-static void uncore_event_init_cpu(int cpu)
+static int uncore_event_cpu_online(unsigned int cpu)
{
int target;
@@ -1177,50 +1203,15 @@ static void uncore_event_init_cpu(int cpu)
*/
target = cpumask_any_and(&uncore_cpu_mask, topology_core_cpumask(cpu));
if (target < nr_cpu_ids)
- return;
+ return 0;
cpumask_set_cpu(cpu, &uncore_cpu_mask);
uncore_change_context(uncore_msr_uncores, -1, cpu);
uncore_change_context(uncore_pci_uncores, -1, cpu);
+ return 0;
}
-static int uncore_cpu_notifier(struct notifier_block *self,
- unsigned long action, void *hcpu)
-{
- unsigned int cpu = (long)hcpu;
-
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_UP_PREPARE:
- return notifier_from_errno(uncore_cpu_prepare(cpu));
-
- case CPU_STARTING:
- uncore_cpu_starting(cpu, false);
- case CPU_DOWN_FAILED:
- uncore_event_init_cpu(cpu);
- break;
-
- case CPU_UP_CANCELED:
- case CPU_DYING:
- uncore_cpu_dying(cpu);
- break;
-
- case CPU_DOWN_PREPARE:
- uncore_event_exit_cpu(cpu);
- break;
- }
- return NOTIFY_OK;
-}
-
-static struct notifier_block uncore_cpu_nb = {
- .notifier_call = uncore_cpu_notifier,
- /*
- * to migrate uncore events, our notifier should be executed
- * before perf core's notifier.
- */
- .priority = CPU_PRI_PERF + 1,
-};
-
static int __init type_pmu_register(struct intel_uncore_type *type)
{
int i, ret;
@@ -1264,41 +1255,6 @@ err:
return ret;
}
-static void __init uncore_cpu_setup(void *dummy)
-{
- uncore_cpu_starting(smp_processor_id(), true);
-}
-
-/* Lazy to avoid allocation of a few bytes for the normal case */
-static __initdata DECLARE_BITMAP(packages, MAX_LOCAL_APIC);
-
-static int __init uncore_cpumask_init(bool msr)
-{
- unsigned int cpu;
-
- for_each_online_cpu(cpu) {
- unsigned int pkg = topology_logical_package_id(cpu);
- int ret;
-
- if (test_and_set_bit(pkg, packages))
- continue;
- /*
- * The first online cpu of each package allocates and takes
- * the refcounts for all other online cpus in that package.
- * If msrs are not enabled no allocation is required.
- */
- if (msr) {
- ret = uncore_cpu_prepare(cpu);
- if (ret)
- return ret;
- }
- uncore_event_init_cpu(cpu);
- smp_call_function_single(cpu, uncore_cpu_setup, NULL, 1);
- }
- __register_cpu_notifier(&uncore_cpu_nb);
- return 0;
-}
-
#define X86_UNCORE_MODEL_MATCH(model, init) \
{ X86_VENDOR_INTEL, 6, model, X86_FEATURE_ANY, (unsigned long)&init }
@@ -1361,30 +1317,32 @@ static const struct intel_uncore_init_fun knl_uncore_init __initconst = {
};
static const struct intel_uncore_init_fun skl_uncore_init __initconst = {
+ .cpu_init = skl_uncore_cpu_init,
.pci_init = skl_uncore_pci_init,
};
static const struct x86_cpu_id intel_uncore_match[] __initconst = {
- X86_UNCORE_MODEL_MATCH(26, nhm_uncore_init), /* Nehalem */
- X86_UNCORE_MODEL_MATCH(30, nhm_uncore_init),
- X86_UNCORE_MODEL_MATCH(37, nhm_uncore_init), /* Westmere */
- X86_UNCORE_MODEL_MATCH(44, nhm_uncore_init),
- X86_UNCORE_MODEL_MATCH(42, snb_uncore_init), /* Sandy Bridge */
- X86_UNCORE_MODEL_MATCH(58, ivb_uncore_init), /* Ivy Bridge */
- X86_UNCORE_MODEL_MATCH(60, hsw_uncore_init), /* Haswell */
- X86_UNCORE_MODEL_MATCH(69, hsw_uncore_init), /* Haswell Celeron */
- X86_UNCORE_MODEL_MATCH(70, hsw_uncore_init), /* Haswell */
- X86_UNCORE_MODEL_MATCH(61, bdw_uncore_init), /* Broadwell */
- X86_UNCORE_MODEL_MATCH(71, bdw_uncore_init), /* Broadwell */
- X86_UNCORE_MODEL_MATCH(45, snbep_uncore_init), /* Sandy Bridge-EP */
- X86_UNCORE_MODEL_MATCH(46, nhmex_uncore_init), /* Nehalem-EX */
- X86_UNCORE_MODEL_MATCH(47, nhmex_uncore_init), /* Westmere-EX aka. Xeon E7 */
- X86_UNCORE_MODEL_MATCH(62, ivbep_uncore_init), /* Ivy Bridge-EP */
- X86_UNCORE_MODEL_MATCH(63, hswep_uncore_init), /* Haswell-EP */
- X86_UNCORE_MODEL_MATCH(79, bdx_uncore_init), /* BDX-EP */
- X86_UNCORE_MODEL_MATCH(86, bdx_uncore_init), /* BDX-DE */
- X86_UNCORE_MODEL_MATCH(87, knl_uncore_init), /* Knights Landing */
- X86_UNCORE_MODEL_MATCH(94, skl_uncore_init), /* SkyLake */
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_NEHALEM_EP, nhm_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_NEHALEM, nhm_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_WESTMERE, nhm_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_WESTMERE_EP, nhm_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_SANDYBRIDGE, snb_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_IVYBRIDGE, ivb_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_HASWELL_CORE, hsw_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_HASWELL_ULT, hsw_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_HASWELL_GT3E, hsw_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_BROADWELL_CORE, bdw_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_BROADWELL_GT3E, bdw_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_SANDYBRIDGE_X, snbep_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_NEHALEM_EX, nhmex_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_WESTMERE_EX, nhmex_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_IVYBRIDGE_X, ivbep_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_HASWELL_X, hswep_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_BROADWELL_X, bdx_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_BROADWELL_XEON_D, bdx_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_XEON_PHI_KNL, knl_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_SKYLAKE_DESKTOP,skl_uncore_init),
+ X86_UNCORE_MODEL_MATCH(INTEL_FAM6_SKYLAKE_MOBILE, skl_uncore_init),
{},
};
@@ -1420,11 +1378,33 @@ static int __init intel_uncore_init(void)
if (cret && pret)
return -ENODEV;
- cpu_notifier_register_begin();
- ret = uncore_cpumask_init(!cret);
- if (ret)
- goto err;
- cpu_notifier_register_done();
+ /*
+ * Install callbacks. Core will call them for each online cpu.
+ *
+ * The first online cpu of each package allocates and takes
+ * the refcounts for all other online cpus in that package.
+ * If msrs are not enabled no allocation is required and
+ * uncore_cpu_prepare() is not called for each online cpu.
+ */
+ if (!cret) {
+ ret = cpuhp_setup_state(CPUHP_PERF_X86_UNCORE_PREP,
+ "PERF_X86_UNCORE_PREP",
+ uncore_cpu_prepare, NULL);
+ if (ret)
+ goto err;
+ } else {
+ cpuhp_setup_state_nocalls(CPUHP_PERF_X86_UNCORE_PREP,
+ "PERF_X86_UNCORE_PREP",
+ uncore_cpu_prepare, NULL);
+ }
+ first_init = 1;
+ cpuhp_setup_state(CPUHP_AP_PERF_X86_UNCORE_STARTING,
+ "AP_PERF_X86_UNCORE_STARTING",
+ uncore_cpu_starting, uncore_cpu_dying);
+ first_init = 0;
+ cpuhp_setup_state(CPUHP_AP_PERF_X86_UNCORE_ONLINE,
+ "AP_PERF_X86_UNCORE_ONLINE",
+ uncore_event_cpu_online, uncore_event_cpu_offline);
return 0;
err:
@@ -1432,17 +1412,16 @@ err:
on_each_cpu_mask(&uncore_cpu_mask, uncore_exit_boxes, NULL, 1);
uncore_types_exit(uncore_msr_uncores);
uncore_pci_exit();
- cpu_notifier_register_done();
return ret;
}
module_init(intel_uncore_init);
static void __exit intel_uncore_exit(void)
{
- cpu_notifier_register_begin();
- __unregister_cpu_notifier(&uncore_cpu_nb);
+ cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_UNCORE_ONLINE);
+ cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_UNCORE_STARTING);
+ cpuhp_remove_state_nocalls(CPUHP_PERF_X86_UNCORE_PREP);
uncore_types_exit(uncore_msr_uncores);
uncore_pci_exit();
- cpu_notifier_register_done();
}
module_exit(intel_uncore_exit);
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 79766b9a3..78b9c23e2 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -1,4 +1,3 @@
-#include <linux/module.h>
#include <linux/slab.h>
#include <linux/pci.h>
#include <asm/apicdef.h>
@@ -15,7 +14,11 @@
#define UNCORE_PMC_IDX_FIXED UNCORE_PMC_IDX_MAX_GENERIC
#define UNCORE_PMC_IDX_MAX (UNCORE_PMC_IDX_FIXED + 1)
+#define UNCORE_PCI_DEV_FULL_DATA(dev, func, type, idx) \
+ ((dev << 24) | (func << 16) | (type << 8) | idx)
#define UNCORE_PCI_DEV_DATA(type, idx) ((type << 8) | idx)
+#define UNCORE_PCI_DEV_DEV(data) ((data >> 24) & 0xff)
+#define UNCORE_PCI_DEV_FUNC(data) ((data >> 16) & 0xff)
#define UNCORE_PCI_DEV_TYPE(data) ((data >> 8) & 0xff)
#define UNCORE_PCI_DEV_IDX(data) (data & 0xff)
#define UNCORE_EXTRA_PCI_DEV 0xff
@@ -360,6 +363,7 @@ int bdw_uncore_pci_init(void);
int skl_uncore_pci_init(void);
void snb_uncore_cpu_init(void);
void nhm_uncore_cpu_init(void);
+void skl_uncore_cpu_init(void);
int snb_pci2phy_map_init(int devid);
/* perf_event_intel_uncore_snbep.c */
diff --git a/arch/x86/events/intel/uncore_snb.c b/arch/x86/events/intel/uncore_snb.c
index 96531d2b8..9d35ec0cb 100644
--- a/arch/x86/events/intel/uncore_snb.c
+++ b/arch/x86/events/intel/uncore_snb.c
@@ -1,4 +1,4 @@
-/* Nehalem/SandBridge/Haswell uncore support */
+/* Nehalem/SandBridge/Haswell/Broadwell/Skylake uncore support */
#include "uncore.h"
/* Uncore IMC PCI IDs */
@@ -9,6 +9,7 @@
#define PCI_DEVICE_ID_INTEL_HSW_U_IMC 0x0a04
#define PCI_DEVICE_ID_INTEL_BDW_IMC 0x1604
#define PCI_DEVICE_ID_INTEL_SKL_IMC 0x191f
+#define PCI_DEVICE_ID_INTEL_SKL_U_IMC 0x190c
/* SNB event control */
#define SNB_UNC_CTL_EV_SEL_MASK 0x000000ff
@@ -64,6 +65,10 @@
#define NHM_UNC_PERFEVTSEL0 0x3c0
#define NHM_UNC_UNCORE_PMC0 0x3b0
+/* SKL uncore global control */
+#define SKL_UNC_PERF_GLOBAL_CTL 0xe01
+#define SKL_UNC_GLOBAL_CTL_CORE_ALL ((1 << 5) - 1)
+
DEFINE_UNCORE_FORMAT_ATTR(event, event, "config:0-7");
DEFINE_UNCORE_FORMAT_ATTR(umask, umask, "config:8-15");
DEFINE_UNCORE_FORMAT_ATTR(edge, edge, "config:18");
@@ -95,6 +100,12 @@ static void snb_uncore_msr_init_box(struct intel_uncore_box *box)
}
}
+static void snb_uncore_msr_enable_box(struct intel_uncore_box *box)
+{
+ wrmsrl(SNB_UNC_PERF_GLOBAL_CTL,
+ SNB_UNC_GLOBAL_CTL_EN | SNB_UNC_GLOBAL_CTL_CORE_ALL);
+}
+
static void snb_uncore_msr_exit_box(struct intel_uncore_box *box)
{
if (box->pmu->pmu_idx == 0)
@@ -122,6 +133,7 @@ static struct attribute_group snb_uncore_format_group = {
static struct intel_uncore_ops snb_uncore_msr_ops = {
.init_box = snb_uncore_msr_init_box,
+ .enable_box = snb_uncore_msr_enable_box,
.exit_box = snb_uncore_msr_exit_box,
.disable_event = snb_uncore_msr_disable_event,
.enable_event = snb_uncore_msr_enable_event,
@@ -179,6 +191,67 @@ void snb_uncore_cpu_init(void)
snb_uncore_cbox.num_boxes = boot_cpu_data.x86_max_cores;
}
+static void skl_uncore_msr_init_box(struct intel_uncore_box *box)
+{
+ if (box->pmu->pmu_idx == 0) {
+ wrmsrl(SKL_UNC_PERF_GLOBAL_CTL,
+ SNB_UNC_GLOBAL_CTL_EN | SKL_UNC_GLOBAL_CTL_CORE_ALL);
+ }
+}
+
+static void skl_uncore_msr_enable_box(struct intel_uncore_box *box)
+{
+ wrmsrl(SKL_UNC_PERF_GLOBAL_CTL,
+ SNB_UNC_GLOBAL_CTL_EN | SKL_UNC_GLOBAL_CTL_CORE_ALL);
+}
+
+static void skl_uncore_msr_exit_box(struct intel_uncore_box *box)
+{
+ if (box->pmu->pmu_idx == 0)
+ wrmsrl(SKL_UNC_PERF_GLOBAL_CTL, 0);
+}
+
+static struct intel_uncore_ops skl_uncore_msr_ops = {
+ .init_box = skl_uncore_msr_init_box,
+ .enable_box = skl_uncore_msr_enable_box,
+ .exit_box = skl_uncore_msr_exit_box,
+ .disable_event = snb_uncore_msr_disable_event,
+ .enable_event = snb_uncore_msr_enable_event,
+ .read_counter = uncore_msr_read_counter,
+};
+
+static struct intel_uncore_type skl_uncore_cbox = {
+ .name = "cbox",
+ .num_counters = 4,
+ .num_boxes = 5,
+ .perf_ctr_bits = 44,
+ .fixed_ctr_bits = 48,
+ .perf_ctr = SNB_UNC_CBO_0_PER_CTR0,
+ .event_ctl = SNB_UNC_CBO_0_PERFEVTSEL0,
+ .fixed_ctr = SNB_UNC_FIXED_CTR,
+ .fixed_ctl = SNB_UNC_FIXED_CTR_CTRL,
+ .single_fixed = 1,
+ .event_mask = SNB_UNC_RAW_EVENT_MASK,
+ .msr_offset = SNB_UNC_CBO_MSR_OFFSET,
+ .ops = &skl_uncore_msr_ops,
+ .format_group = &snb_uncore_format_group,
+ .event_descs = snb_uncore_events,
+};
+
+static struct intel_uncore_type *skl_msr_uncores[] = {
+ &skl_uncore_cbox,
+ &snb_uncore_arb,
+ NULL,
+};
+
+void skl_uncore_cpu_init(void)
+{
+ uncore_msr_uncores = skl_msr_uncores;
+ if (skl_uncore_cbox.num_boxes > boot_cpu_data.x86_max_cores)
+ skl_uncore_cbox.num_boxes = boot_cpu_data.x86_max_cores;
+ snb_uncore_arb.ops = &skl_uncore_msr_ops;
+}
+
enum {
SNB_PCI_UNCORE_IMC,
};
@@ -544,6 +617,11 @@ static const struct pci_device_id skl_uncore_pci_ids[] = {
PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_SKL_IMC),
.driver_data = UNCORE_PCI_DEV_DATA(SNB_PCI_UNCORE_IMC, 0),
},
+ { /* IMC */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_SKL_U_IMC),
+ .driver_data = UNCORE_PCI_DEV_DATA(SNB_PCI_UNCORE_IMC, 0),
+ },
+
{ /* end: all zeroes */ },
};
@@ -587,6 +665,7 @@ static const struct imc_uncore_pci_dev desktop_imc_pci_ids[] = {
IMC_DEV(HSW_U_IMC, &hsw_uncore_pci_driver), /* 4th Gen Core ULT Mobile Processor */
IMC_DEV(BDW_IMC, &bdw_uncore_pci_driver), /* 5th Gen Core U */
IMC_DEV(SKL_IMC, &skl_uncore_pci_driver), /* 6th Gen Core */
+ IMC_DEV(SKL_U_IMC, &skl_uncore_pci_driver), /* 6th Gen Core U */
{ /* end marker */ }
};
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index bd136ac14..8aee83bcf 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -2164,21 +2164,101 @@ static struct intel_uncore_type *knl_pci_uncores[] = {
*/
static const struct pci_device_id knl_uncore_pci_ids[] = {
- { /* MC UClk */
+ { /* MC0 UClk */
PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7841),
- .driver_data = UNCORE_PCI_DEV_DATA(KNL_PCI_UNCORE_MC_UCLK, 0),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(10, 0, KNL_PCI_UNCORE_MC_UCLK, 0),
},
- { /* MC DClk Channel */
+ { /* MC1 UClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7841),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(11, 0, KNL_PCI_UNCORE_MC_UCLK, 1),
+ },
+ { /* MC0 DClk CH 0 */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7843),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(8, 2, KNL_PCI_UNCORE_MC_DCLK, 0),
+ },
+ { /* MC0 DClk CH 1 */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7843),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(8, 3, KNL_PCI_UNCORE_MC_DCLK, 1),
+ },
+ { /* MC0 DClk CH 2 */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7843),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(8, 4, KNL_PCI_UNCORE_MC_DCLK, 2),
+ },
+ { /* MC1 DClk CH 0 */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7843),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(9, 2, KNL_PCI_UNCORE_MC_DCLK, 3),
+ },
+ { /* MC1 DClk CH 1 */
PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7843),
- .driver_data = UNCORE_PCI_DEV_DATA(KNL_PCI_UNCORE_MC_DCLK, 0),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(9, 3, KNL_PCI_UNCORE_MC_DCLK, 4),
+ },
+ { /* MC1 DClk CH 2 */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7843),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(9, 4, KNL_PCI_UNCORE_MC_DCLK, 5),
+ },
+ { /* EDC0 UClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7833),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(15, 0, KNL_PCI_UNCORE_EDC_UCLK, 0),
+ },
+ { /* EDC1 UClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7833),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(16, 0, KNL_PCI_UNCORE_EDC_UCLK, 1),
+ },
+ { /* EDC2 UClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7833),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(17, 0, KNL_PCI_UNCORE_EDC_UCLK, 2),
+ },
+ { /* EDC3 UClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7833),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(18, 0, KNL_PCI_UNCORE_EDC_UCLK, 3),
},
- { /* EDC UClk */
+ { /* EDC4 UClk */
PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7833),
- .driver_data = UNCORE_PCI_DEV_DATA(KNL_PCI_UNCORE_EDC_UCLK, 0),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(19, 0, KNL_PCI_UNCORE_EDC_UCLK, 4),
+ },
+ { /* EDC5 UClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7833),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(20, 0, KNL_PCI_UNCORE_EDC_UCLK, 5),
+ },
+ { /* EDC6 UClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7833),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(21, 0, KNL_PCI_UNCORE_EDC_UCLK, 6),
+ },
+ { /* EDC7 UClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7833),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(22, 0, KNL_PCI_UNCORE_EDC_UCLK, 7),
+ },
+ { /* EDC0 EClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7835),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(24, 2, KNL_PCI_UNCORE_EDC_ECLK, 0),
+ },
+ { /* EDC1 EClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7835),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(25, 2, KNL_PCI_UNCORE_EDC_ECLK, 1),
+ },
+ { /* EDC2 EClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7835),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(26, 2, KNL_PCI_UNCORE_EDC_ECLK, 2),
+ },
+ { /* EDC3 EClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7835),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(27, 2, KNL_PCI_UNCORE_EDC_ECLK, 3),
+ },
+ { /* EDC4 EClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7835),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(28, 2, KNL_PCI_UNCORE_EDC_ECLK, 4),
+ },
+ { /* EDC5 EClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7835),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(29, 2, KNL_PCI_UNCORE_EDC_ECLK, 5),
+ },
+ { /* EDC6 EClk */
+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7835),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(30, 2, KNL_PCI_UNCORE_EDC_ECLK, 6),
},
- { /* EDC EClk */
+ { /* EDC7 EClk */
PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7835),
- .driver_data = UNCORE_PCI_DEV_DATA(KNL_PCI_UNCORE_EDC_ECLK, 0),
+ .driver_data = UNCORE_PCI_DEV_FULL_DATA(31, 2, KNL_PCI_UNCORE_EDC_ECLK, 7),
},
{ /* M2PCIe */
PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7817),
diff --git a/arch/x86/events/msr.c b/arch/x86/events/msr.c
index 85ef3c2e8..4bb3ec69e 100644
--- a/arch/x86/events/msr.c
+++ b/arch/x86/events/msr.c
@@ -1,4 +1,5 @@
#include <linux/perf_event.h>
+#include <asm/intel-family.h>
enum perf_msr_id {
PERF_MSR_TSC = 0,
@@ -34,39 +35,43 @@ static bool test_intel(int idx)
return false;
switch (boot_cpu_data.x86_model) {
- case 30: /* 45nm Nehalem */
- case 26: /* 45nm Nehalem-EP */
- case 46: /* 45nm Nehalem-EX */
-
- case 37: /* 32nm Westmere */
- case 44: /* 32nm Westmere-EP */
- case 47: /* 32nm Westmere-EX */
-
- case 42: /* 32nm SandyBridge */
- case 45: /* 32nm SandyBridge-E/EN/EP */
-
- case 58: /* 22nm IvyBridge */
- case 62: /* 22nm IvyBridge-EP/EX */
-
- case 60: /* 22nm Haswell Core */
- case 63: /* 22nm Haswell Server */
- case 69: /* 22nm Haswell ULT */
- case 70: /* 22nm Haswell + GT3e (Intel Iris Pro graphics) */
-
- case 61: /* 14nm Broadwell Core-M */
- case 86: /* 14nm Broadwell Xeon D */
- case 71: /* 14nm Broadwell + GT3e (Intel Iris Pro graphics) */
- case 79: /* 14nm Broadwell Server */
-
- case 55: /* 22nm Atom "Silvermont" */
- case 77: /* 22nm Atom "Silvermont Avoton/Rangely" */
- case 76: /* 14nm Atom "Airmont" */
+ case INTEL_FAM6_NEHALEM:
+ case INTEL_FAM6_NEHALEM_G:
+ case INTEL_FAM6_NEHALEM_EP:
+ case INTEL_FAM6_NEHALEM_EX:
+
+ case INTEL_FAM6_WESTMERE:
+ case INTEL_FAM6_WESTMERE_EP:
+ case INTEL_FAM6_WESTMERE_EX:
+
+ case INTEL_FAM6_SANDYBRIDGE:
+ case INTEL_FAM6_SANDYBRIDGE_X:
+
+ case INTEL_FAM6_IVYBRIDGE:
+ case INTEL_FAM6_IVYBRIDGE_X:
+
+ case INTEL_FAM6_HASWELL_CORE:
+ case INTEL_FAM6_HASWELL_X:
+ case INTEL_FAM6_HASWELL_ULT:
+ case INTEL_FAM6_HASWELL_GT3E:
+
+ case INTEL_FAM6_BROADWELL_CORE:
+ case INTEL_FAM6_BROADWELL_XEON_D:
+ case INTEL_FAM6_BROADWELL_GT3E:
+ case INTEL_FAM6_BROADWELL_X:
+
+ case INTEL_FAM6_ATOM_SILVERMONT1:
+ case INTEL_FAM6_ATOM_SILVERMONT2:
+ case INTEL_FAM6_ATOM_AIRMONT:
if (idx == PERF_MSR_SMI)
return true;
break;
- case 78: /* 14nm Skylake Mobile */
- case 94: /* 14nm Skylake Desktop */
+ case INTEL_FAM6_SKYLAKE_MOBILE:
+ case INTEL_FAM6_SKYLAKE_DESKTOP:
+ case INTEL_FAM6_SKYLAKE_X:
+ case INTEL_FAM6_KABYLAKE_MOBILE:
+ case INTEL_FAM6_KABYLAKE_DESKTOP:
if (idx == PERF_MSR_SMI || idx == PERF_MSR_PPERF)
return true;
break;
diff --git a/arch/x86/events/perf_event.h b/arch/x86/events/perf_event.h
index 8bd764df8..8c4a47706 100644
--- a/arch/x86/events/perf_event.h
+++ b/arch/x86/events/perf_event.h
@@ -668,6 +668,14 @@ static struct perf_pmu_events_attr event_attr_##v = { \
.event_str = str, \
};
+#define EVENT_ATTR_STR_HT(_name, v, noht, ht) \
+static struct perf_pmu_events_ht_attr event_attr_##v = { \
+ .attr = __ATTR(_name, 0444, events_ht_sysfs_show, NULL),\
+ .id = 0, \
+ .event_str_noht = noht, \
+ .event_str_ht = ht, \
+}
+
extern struct x86_pmu x86_pmu __read_mostly;
static inline bool x86_pmu_has_lbr_callstack(void)
@@ -803,6 +811,8 @@ struct attribute **merge_attr(struct attribute **a, struct attribute **b);
ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
char *page);
+ssize_t events_ht_sysfs_show(struct device *dev, struct device_attribute *attr,
+ char *page);
#ifdef CONFIG_CPU_SUP_AMD
@@ -892,6 +902,8 @@ void intel_ds_init(void);
void intel_pmu_lbr_sched_task(struct perf_event_context *ctx, bool sched_in);
+u64 lbr_from_signext_quirk_wr(u64 val);
+
void intel_pmu_lbr_reset(void);
void intel_pmu_lbr_enable(struct perf_event *event);
generated by cgit v1.2.3-54-g00ecf (git 2.45.2) at 2024-11-16 12:02:40 +0000