Merge tag 'perf-core-2024-01-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull performance events updates from Ingo Molnar:

- Add branch stack counters ABI extension to better capture the growing
amount of information the PMU exposes via branch stack sampling.
There's matching tooling support.

- Fix race when creating the nr_addr_filters sysfs file

- Add Intel Sierra Forest and Grand Ridge intel/cstate PMU support

- Add Intel Granite Rapids, Sierra Forest and Grand Ridge uncore PMU
support

- Misc cleanups & fixes

* tag 'perf-core-2024-01-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel/uncore: Factor out topology_gidnid_map()
perf/x86/intel/uncore: Fix NULL pointer dereference issue in upi_fill_topology()
perf/x86/amd: Reject branch stack for IBS events
perf/x86/intel/uncore: Support Sierra Forest and Grand Ridge
perf/x86/intel/uncore: Support IIO free-running counters on GNR
perf/x86/intel/uncore: Support Granite Rapids
perf/x86/uncore: Use u64 to replace unsigned for the uncore offsets array
perf/x86/intel/uncore: Generic uncore_get_uncores and MMIO format of SPR
perf: Fix the nr_addr_filters fix
perf/x86/intel/cstate: Add Grand Ridge support
perf/x86/intel/cstate: Add Sierra Forest support
x86/smp: Export symbol cpu_clustergroup_mask()
perf/x86/intel/cstate: Cleanup duplicate attr_groups
perf/core: Fix narrow startup race when creating the perf nr_addr_filters sysfs file
perf/x86/intel: Support branch counters logging
perf/x86/intel: Reorganize attrs and is_visible
perf: Add branch_sample_call_stack
perf/x86: Add PERF_X86_EVENT_NEEDS_BRANCH_STACK flag
perf: Add branch stack counters

Documentation/ABI/testing/sysfs-bus-event_source-devices-caps

···
 		Example output in powerpc:
 		grep . /sys/bus/event_source/devices/cpu/caps/*
 		/sys/bus/event_source/devices/cpu/caps/pmu_name:POWER9
+
+		The "branch_counter_nr" in the supported platform exposes the
+		maximum number of counters which can be shown in the u64 counters
+		of PERF_SAMPLE_BRANCH_COUNTERS, while the "branch_counter_width"
+		exposes the width of each counter. Both of them can be used by
+		the perf tool to parse the logged counters in each branch.

arch/powerpc/perf/core-book3s.c

···
 			struct cpu_hw_events *cpuhw;
 			cpuhw = this_cpu_ptr(&cpu_hw_events);
 			power_pmu_bhrb_read(event, cpuhw);
-			perf_sample_save_brstack(&data, event, &cpuhw->bhrb_stack);
+			perf_sample_save_brstack(&data, event, &cpuhw->bhrb_stack, NULL);
 		}
 
 		if (event->attr.sample_type & PERF_SAMPLE_DATA_SRC &&

arch/x86/events/amd/core.c

···
 			continue;
 
 		if (has_branch_stack(event))
-			perf_sample_save_brstack(&data, event, &cpuc->lbr_stack);
+			perf_sample_save_brstack(&data, event, &cpuc->lbr_stack, NULL);
 
 		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);

arch/x86/events/amd/ibs.c

···
 	if (config & ~perf_ibs->config_mask)
 		return -EINVAL;
 
+	if (has_branch_stack(event))
+		return -EOPNOTSUPP;
+
 	ret = validate_group(event);
 	if (ret)
 		return ret;

arch/x86/events/core.c

···
 		}
 	}
 
-	if (event->attr.branch_sample_type & PERF_SAMPLE_BRANCH_CALL_STACK)
+	if (branch_sample_call_stack(event))
 		event->attach_state |= PERF_ATTACH_TASK_DATA;
 
 	/*
···
 		perf_sample_data_init(&data, 0, event->hw.last_period);
 
 		if (has_branch_stack(event))
-			perf_sample_save_brstack(&data, event, &cpuc->lbr_stack);
+			perf_sample_save_brstack(&data, event, &cpuc->lbr_stack, NULL);
 
 		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);

arch/x86/events/intel/core.c

···
 		perf_report_aux_output_id(event, idx);
 }
 
+static __always_inline bool intel_pmu_needs_branch_stack(struct perf_event *event)
+{
+	return event->hw.flags & PERF_X86_EVENT_NEEDS_BRANCH_STACK;
+}
+
 static void intel_pmu_del_event(struct perf_event *event)
 {
-	if (needs_branch_stack(event))
+	if (intel_pmu_needs_branch_stack(event))
 		intel_pmu_lbr_del(event);
 	if (event->attr.precise_ip)
 		intel_pmu_pebs_del(event);
···
 
 static void intel_pmu_enable_event(struct perf_event *event)
 {
+	u64 enable_mask = ARCH_PERFMON_EVENTSEL_ENABLE;
 	struct hw_perf_event *hwc = &event->hw;
 	int idx = hwc->idx;
 
···
 
 	switch (idx) {
 	case 0 ... INTEL_PMC_IDX_FIXED - 1:
+		if (branch_sample_counters(event))
+			enable_mask |= ARCH_PERFMON_EVENTSEL_BR_CNTR;
 		intel_set_masks(event, idx);
-		__x86_pmu_enable_event(hwc, ARCH_PERFMON_EVENTSEL_ENABLE);
+		__x86_pmu_enable_event(hwc, enable_mask);
 		break;
 	case INTEL_PMC_IDX_FIXED ... INTEL_PMC_IDX_FIXED_BTS - 1:
 	case INTEL_PMC_IDX_METRIC_BASE ... INTEL_PMC_IDX_METRIC_END:
···
 {
 	if (event->attr.precise_ip)
 		intel_pmu_pebs_add(event);
-	if (needs_branch_stack(event))
+	if (intel_pmu_needs_branch_stack(event))
 		intel_pmu_lbr_add(event);
 }
 
···
 		perf_sample_data_init(&data, 0, event->hw.last_period);
 
 		if (has_branch_stack(event))
-			perf_sample_save_brstack(&data, event, &cpuc->lbr_stack);
+			intel_pmu_lbr_save_brstack(&data, cpuc, event);
 
 		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);
···
 	if (cpuc->excl_cntrs)
 		return intel_get_excl_constraints(cpuc, event, idx, c2);
 
+	/* Not all counters support the branch counter feature. */
+	if (branch_sample_counters(event)) {
+		c2 = dyn_constraint(cpuc, c2, idx);
+		c2->idxmsk64 &= x86_pmu.lbr_counters;
+		c2->weight = hweight64(c2->idxmsk64);
+	}
+
 	return c2;
 }
 
···
 			x86_pmu.pebs_aliases(event);
 	}
 
-	if (needs_branch_stack(event)) {
+	if (needs_branch_stack(event) && is_sampling_event(event))
+		event->hw.flags  |= PERF_X86_EVENT_NEEDS_BRANCH_STACK;
+
+	if (branch_sample_counters(event)) {
+		struct perf_event *leader, *sibling;
+		int num = 0;
+
+		if (!(x86_pmu.flags & PMU_FL_BR_CNTR) ||
+		    (event->attr.config & ~INTEL_ARCH_EVENT_MASK))
+			return -EINVAL;
+
+		/*
+		 * The branch counter logging is not supported in the call stack
+		 * mode yet, since we cannot simply flush the LBR during e.g.,
+		 * multiplexing. Also, there is no obvious usage with the call
+		 * stack mode. Simply forbids it for now.
+		 *
+		 * If any events in the group enable the branch counter logging
+		 * feature, the group is treated as a branch counter logging
+		 * group, which requires the extra space to store the counters.
+		 */
+		leader = event->group_leader;
+		if (branch_sample_call_stack(leader))
+			return -EINVAL;
+		if (branch_sample_counters(leader))
+			num++;
+		leader->hw.flags |= PERF_X86_EVENT_BRANCH_COUNTERS;
+
+		for_each_sibling_event(sibling, leader) {
+			if (branch_sample_call_stack(sibling))
+				return -EINVAL;
+			if (branch_sample_counters(sibling))
+				num++;
+		}
+
+		if (num > fls(x86_pmu.lbr_counters))
+			return -EINVAL;
+		/*
+		 * Only applying the PERF_SAMPLE_BRANCH_COUNTERS doesn't
+		 * require any branch stack setup.
+		 * Clear the bit to avoid unnecessary branch stack setup.
+		 */
+		if (0 == (event->attr.branch_sample_type &
+			  ~(PERF_SAMPLE_BRANCH_PLM_ALL |
+			    PERF_SAMPLE_BRANCH_COUNTERS)))
+			event->hw.flags  &= ~PERF_X86_EVENT_NEEDS_BRANCH_STACK;
+
+		/*
+		 * Force the leader to be a LBR event. So LBRs can be reset
+		 * with the leader event. See intel_pmu_lbr_del() for details.
+		 */
+		if (!intel_pmu_needs_branch_stack(leader))
+			return -EINVAL;
+	}
+
+	if (intel_pmu_needs_branch_stack(event)) {
 		ret = intel_pmu_setup_lbr_filter(event);
 		if (ret)
 			return ret;
···
 	 */
 	if (event->attr.precise_ip == 3) {
 		/* Force instruction:ppp on PMC0, 1 and Fixed counter 0 */
-		if (constraint_match(&fixed0_constraint, event->hw.config))
-			return &fixed0_counter0_1_constraint;
+		if (constraint_match(&fixed0_constraint, event->hw.config)) {
+			/* The fixed counter 0 doesn't support LBR event logging. */
+			if (branch_sample_counters(event))
+				return &counter0_1_constraint;
+			else
+				return &fixed0_counter0_1_constraint;
+		}
 
 		switch (c->idxmsk64 & 0x3ull) {
 		case 0x1:
···
 			goto err;
 	}
 
-	if (x86_pmu.flags & (PMU_FL_EXCL_CNTRS | PMU_FL_TFA)) {
+	if (x86_pmu.flags & (PMU_FL_EXCL_CNTRS | PMU_FL_TFA | PMU_FL_BR_CNTR)) {
 		size_t sz = X86_PMC_IDX_MAX * sizeof(struct event_constraint);
 
 		cpuc->constraint_list = kzalloc_node(sz, GFP_KERNEL, cpu_to_node(cpu));
···
 
 static DEVICE_ATTR_RO(branches);
 
+static ssize_t branch_counter_nr_show(struct device *cdev,
+				      struct device_attribute *attr,
+				      char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%d\n", fls(x86_pmu.lbr_counters));
+}
+
+static DEVICE_ATTR_RO(branch_counter_nr);
+
+static ssize_t branch_counter_width_show(struct device *cdev,
+					 struct device_attribute *attr,
+					 char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%d\n", LBR_INFO_BR_CNTR_BITS);
+}
+
+static DEVICE_ATTR_RO(branch_counter_width);
+
 static struct attribute *lbr_attrs[] = {
 	&dev_attr_branches.attr,
+	&dev_attr_branch_counter_nr.attr,
+	&dev_attr_branch_counter_width.attr,
 	NULL
 };
+
+static umode_t
+lbr_is_visible(struct kobject *kobj, struct attribute *attr, int i)
+{
+	/* branches */
+	if (i == 0)
+		return x86_pmu.lbr_nr ? attr->mode : 0;
+
+	return (x86_pmu.flags & PMU_FL_BR_CNTR) ? attr->mode : 0;
+}
 
 static char pmu_name_str[30];
 
···
 };
 
 static umode_t
+default_is_visible(struct kobject *kobj, struct attribute *attr, int i)
+{
+	if (attr == &dev_attr_allow_tsx_force_abort.attr)
+		return x86_pmu.flags & PMU_FL_TFA ? attr->mode : 0;
+
+	return attr->mode;
+}
+
+static umode_t
 tsx_is_visible(struct kobject *kobj, struct attribute *attr, int i)
 {
 	return boot_cpu_has(X86_FEATURE_RTM) ? attr->mode : 0;
···
 }
 
 static umode_t
-lbr_is_visible(struct kobject *kobj, struct attribute *attr, int i)
-{
-	return x86_pmu.lbr_nr ? attr->mode : 0;
-}
-
-static umode_t
 exra_is_visible(struct kobject *kobj, struct attribute *attr, int i)
 {
 	return x86_pmu.version >= 2 ? attr->mode : 0;
-}
-
-static umode_t
-default_is_visible(struct kobject *kobj, struct attribute *attr, int i)
-{
-	if (attr == &dev_attr_allow_tsx_force_abort.attr)
-		return x86_pmu.flags & PMU_FL_TFA ? attr->mode : 0;
-
-	return attr->mode;
 }
 
 static struct attribute_group group_events_td  = {

arch/x86/events/intel/cstate.c

···
  *	MSR_CORE_C1_RES: CORE C1 Residency Counter
  *			 perf code: 0x00
  *			 Available model: SLM,AMT,GLM,CNL,ICX,TNT,ADL,RPL
- *					  MTL
+ *					  MTL,SRF,GRR
  *			 Scope: Core (each processor core has a MSR)
  *	MSR_CORE_C3_RESIDENCY: CORE C3 Residency Counter
  *			       perf code: 0x01
···
  *			       perf code: 0x02
  *			       Available model: SLM,AMT,NHM,WSM,SNB,IVB,HSW,BDW,
  *						SKL,KNL,GLM,CNL,KBL,CML,ICL,ICX,
- *						TGL,TNT,RKL,ADL,RPL,SPR,MTL
+ *						TGL,TNT,RKL,ADL,RPL,SPR,MTL,SRF,
+ *						GRR
  *			       Scope: Core
  *	MSR_CORE_C7_RESIDENCY: CORE C7 Residency Counter
  *			       perf code: 0x03
···
  *			       perf code: 0x02
  *			       Available model: SLM,AMT,NHM,WSM,SNB,IVB,HSW,BDW,
  *						SKL,KNL,GLM,CNL,KBL,CML,ICL,ICX,
- *						TGL,TNT,RKL,ADL,RPL,SPR,MTL
+ *						TGL,TNT,RKL,ADL,RPL,SPR,MTL,SRF
  *			       Scope: Package (physical package)
  *	MSR_PKG_C7_RESIDENCY:  Package C7 Residency Counter.
  *			       perf code: 0x03
···
  *			       Available model: HSW ULT,KBL,GLM,CNL,CML,ICL,TGL,
  *						TNT,RKL,ADL,RPL,MTL
  *			       Scope: Package (physical package)
+ *	MSR_MODULE_C6_RES_MS:  Module C6 Residency Counter.
+ *			       perf code: 0x00
+ *			       Available model: SRF,GRR
+ *			       Scope: A cluster of cores shared L2 cache
  *
  */
 
···
 struct cstate_model {
 	unsigned long		core_events;
 	unsigned long		pkg_events;
+	unsigned long		module_events;
 	unsigned long		quirks;
 };
 
···
  * "events" group (with empty attrs) before updating
  * it with detected events.
  */
-static struct attribute_group core_events_attr_group = {
+static struct attribute_group cstate_events_attr_group = {
 	.name = "events",
 	.attrs = attrs_empty,
 };
 
-DEFINE_CSTATE_FORMAT_ATTR(core_event, event, "config:0-63");
-static struct attribute *core_format_attrs[] = {
-	&format_attr_core_event.attr,
+DEFINE_CSTATE_FORMAT_ATTR(cstate_event, event, "config:0-63");
+static struct attribute *cstate_format_attrs[] = {
+	&format_attr_cstate_event.attr,
 	NULL,
 };
 
-static struct attribute_group core_format_attr_group = {
+static struct attribute_group cstate_format_attr_group = {
 	.name = "format",
-	.attrs = core_format_attrs,
+	.attrs = cstate_format_attrs,
 };
 
 static cpumask_t cstate_core_cpu_mask;
···
 	.attrs = cstate_cpumask_attrs,
 };
 
-static const struct attribute_group *core_attr_groups[] = {
-	&core_events_attr_group,
-	&core_format_attr_group,
+static const struct attribute_group *cstate_attr_groups[] = {
+	&cstate_events_attr_group,
+	&cstate_format_attr_group,
 	&cpumask_attr_group,
 	NULL,
 };
···
 	[PERF_CSTATE_PKG_C10_RES] = { MSR_PKG_C10_RESIDENCY,	&group_cstate_pkg_c10,	test_msr },
 };
 
-static struct attribute_group pkg_events_attr_group = {
-	.name = "events",
-	.attrs = attrs_empty,
-};
-
-DEFINE_CSTATE_FORMAT_ATTR(pkg_event, event, "config:0-63");
-static struct attribute *pkg_format_attrs[] = {
-	&format_attr_pkg_event.attr,
-	NULL,
-};
-static struct attribute_group pkg_format_attr_group = {
-	.name = "format",
-	.attrs = pkg_format_attrs,
-};
-
 static cpumask_t cstate_pkg_cpu_mask;
 
-static const struct attribute_group *pkg_attr_groups[] = {
-	&pkg_events_attr_group,
-	&pkg_format_attr_group,
-	&cpumask_attr_group,
-	NULL,
+/* cstate_module PMU */
+static struct pmu cstate_module_pmu;
+static bool has_cstate_module;
+
+enum perf_cstate_module_events {
+	PERF_CSTATE_MODULE_C6_RES = 0,
+
+	PERF_CSTATE_MODULE_EVENT_MAX,
 };
+
+PMU_EVENT_ATTR_STRING(c6-residency, attr_cstate_module_c6, "event=0x00");
+
+static unsigned long module_msr_mask;
+
+PMU_EVENT_GROUP(events, cstate_module_c6);
+
+static struct perf_msr module_msr[] = {
+	[PERF_CSTATE_MODULE_C6_RES]  = { MSR_MODULE_C6_RES_MS,	&group_cstate_module_c6,	test_msr },
+};
+
+static cpumask_t cstate_module_cpu_mask;
 
 static ssize_t cstate_get_attr_cpumask(struct device *dev,
 				       struct device_attribute *attr,
···
 		return cpumap_print_to_pagebuf(true, buf, &cstate_core_cpu_mask);
 	else if (pmu == &cstate_pkg_pmu)
 		return cpumap_print_to_pagebuf(true, buf, &cstate_pkg_cpu_mask);
+	else if (pmu == &cstate_module_pmu)
+		return cpumap_print_to_pagebuf(true, buf, &cstate_module_cpu_mask);
 	else
 		return 0;
 }
···
 		event->hw.event_base = pkg_msr[cfg].msr;
 		cpu = cpumask_any_and(&cstate_pkg_cpu_mask,
 				      topology_die_cpumask(event->cpu));
+	} else if (event->pmu == &cstate_module_pmu) {
+		if (cfg >= PERF_CSTATE_MODULE_EVENT_MAX)
+			return -EINVAL;
+		cfg = array_index_nospec((unsigned long)cfg, PERF_CSTATE_MODULE_EVENT_MAX);
+		if (!(module_msr_mask & (1 << cfg)))
+			return -EINVAL;
+		event->hw.event_base = module_msr[cfg].msr;
+		cpu = cpumask_any_and(&cstate_module_cpu_mask,
+				      topology_cluster_cpumask(event->cpu));
 	} else {
 		return -ENOENT;
 	}
···
 			perf_pmu_migrate_context(&cstate_pkg_pmu, cpu, target);
 		}
 	}
+
+	if (has_cstate_module &&
+	    cpumask_test_and_clear_cpu(cpu, &cstate_module_cpu_mask)) {
+
+		target = cpumask_any_but(topology_cluster_cpumask(cpu), cpu);
+		/* Migrate events if there is a valid target */
+		if (target < nr_cpu_ids) {
+			cpumask_set_cpu(target, &cstate_module_cpu_mask);
+			perf_pmu_migrate_context(&cstate_module_pmu, cpu, target);
+		}
+	}
 	return 0;
 }
 
···
 	if (has_cstate_pkg && target >= nr_cpu_ids)
 		cpumask_set_cpu(cpu, &cstate_pkg_cpu_mask);
 
+	/*
+	 * If this is the first online thread of that cluster, set it
+	 * in the cluster cpu mask as the designated reader.
+	 */
+	target = cpumask_any_and(&cstate_module_cpu_mask,
+				 topology_cluster_cpumask(cpu));
+	if (has_cstate_module && target >= nr_cpu_ids)
+		cpumask_set_cpu(cpu, &cstate_module_cpu_mask);
+
 	return 0;
 }
 
···
 	NULL,
 };
 
+static const struct attribute_group *module_attr_update[] = {
+	&group_cstate_module_c6,
+	NULL
+};
+
 static struct pmu cstate_core_pmu = {
-	.attr_groups	= core_attr_groups,
+	.attr_groups	= cstate_attr_groups,
 	.attr_update	= core_attr_update,
 	.name		= "cstate_core",
 	.task_ctx_nr	= perf_invalid_context,
···
 };
 
 static struct pmu cstate_pkg_pmu = {
-	.attr_groups	= pkg_attr_groups,
+	.attr_groups	= cstate_attr_groups,
 	.attr_update	= pkg_attr_update,
 	.name		= "cstate_pkg",
+	.task_ctx_nr	= perf_invalid_context,
+	.event_init	= cstate_pmu_event_init,
+	.add		= cstate_pmu_event_add,
+	.del		= cstate_pmu_event_del,
+	.start		= cstate_pmu_event_start,
+	.stop		= cstate_pmu_event_stop,
+	.read		= cstate_pmu_event_update,
+	.capabilities	= PERF_PMU_CAP_NO_INTERRUPT | PERF_PMU_CAP_NO_EXCLUDE,
+	.module		= THIS_MODULE,
+};
+
+static struct pmu cstate_module_pmu = {
+	.attr_groups	= cstate_attr_groups,
+	.attr_update	= module_attr_update,
+	.name		= "cstate_module",
 	.task_ctx_nr	= perf_invalid_context,
 	.event_init	= cstate_pmu_event_init,
 	.add		= cstate_pmu_event_add,
···
 				  BIT(PERF_CSTATE_PKG_C10_RES),
 };
 
+static const struct cstate_model grr_cstates __initconst = {
+	.core_events		= BIT(PERF_CSTATE_CORE_C1_RES) |
+				  BIT(PERF_CSTATE_CORE_C6_RES),
+
+	.module_events		= BIT(PERF_CSTATE_MODULE_C6_RES),
+};
+
+static const struct cstate_model srf_cstates __initconst = {
+	.core_events		= BIT(PERF_CSTATE_CORE_C1_RES) |
+				  BIT(PERF_CSTATE_CORE_C6_RES),
+
+	.pkg_events		= BIT(PERF_CSTATE_PKG_C6_RES),
+
+	.module_events		= BIT(PERF_CSTATE_MODULE_C6_RES),
+};
+
 
 static const struct x86_cpu_id intel_cstates_match[] __initconst = {
 	X86_MATCH_INTEL_FAM6_MODEL(NEHALEM,		&nhm_cstates),
···
 	X86_MATCH_INTEL_FAM6_MODEL(ATOM_TREMONT,	&glm_cstates),
 	X86_MATCH_INTEL_FAM6_MODEL(ATOM_TREMONT_L,	&glm_cstates),
 	X86_MATCH_INTEL_FAM6_MODEL(ATOM_GRACEMONT,	&adl_cstates),
+	X86_MATCH_INTEL_FAM6_MODEL(ATOM_CRESTMONT_X,	&srf_cstates),
+	X86_MATCH_INTEL_FAM6_MODEL(ATOM_CRESTMONT,	&grr_cstates),
 
 	X86_MATCH_INTEL_FAM6_MODEL(ICELAKE_L,		&icl_cstates),
 	X86_MATCH_INTEL_FAM6_MODEL(ICELAKE,		&icl_cstates),
···
 	pkg_msr_mask = perf_msr_probe(pkg_msr, PERF_CSTATE_PKG_EVENT_MAX,
 				      true, (void *) &cm->pkg_events);
 
+	module_msr_mask = perf_msr_probe(module_msr, PERF_CSTATE_MODULE_EVENT_MAX,
+				      true, (void *) &cm->module_events);
+
 	has_cstate_core = !!core_msr_mask;
 	has_cstate_pkg  = !!pkg_msr_mask;
+	has_cstate_module  = !!module_msr_mask;
 
-	return (has_cstate_core || has_cstate_pkg) ? 0 : -ENODEV;
+	return (has_cstate_core || has_cstate_pkg || has_cstate_module) ? 0 : -ENODEV;
 }
 
 static inline void cstate_cleanup(void)
···
 
 	if (has_cstate_pkg)
 		perf_pmu_unregister(&cstate_pkg_pmu);
+
+	if (has_cstate_module)
+		perf_pmu_unregister(&cstate_module_pmu);
 }
 
 static int __init cstate_init(void)
···
 		if (err) {
 			has_cstate_pkg = false;
 			pr_info("Failed to register cstate pkg pmu\n");
+			cstate_cleanup();
+			return err;
+		}
+	}
+
+	if (has_cstate_module) {
+		err = perf_pmu_register(&cstate_module_pmu, cstate_module_pmu.name, -1);
+		if (err) {
+			has_cstate_module = false;
+			pr_info("Failed to register cstate cluster pmu\n");
 			cstate_cleanup();
 			return err;
 		}

arch/x86/events/intel/ds.c

···
 		setup_pebs_time(event, data, pebs->tsc);
 
 	if (has_branch_stack(event))
-		perf_sample_save_brstack(data, event, &cpuc->lbr_stack);
+		perf_sample_save_brstack(data, event, &cpuc->lbr_stack, NULL);
 }
 
 static void adaptive_pebs_save_regs(struct pt_regs *regs,
···
 
 		if (has_branch_stack(event)) {
 			intel_pmu_store_pebs_lbrs(lbr);
-			perf_sample_save_brstack(data, event, &cpuc->lbr_stack);
+			intel_pmu_lbr_save_brstack(data, cpuc, event);
 		}
 	}
 

arch/x86/events/intel/lbr.c

···
 	WARN_ON_ONCE(cpuc->lbr_users < 0);
 	WARN_ON_ONCE(cpuc->lbr_pebs_users < 0);
 	perf_sched_cb_dec(event->pmu);
+
+	/*
+	 * The logged occurrences information is only valid for the
+	 * current LBR group. If another LBR group is scheduled in
+	 * later, the information from the stale LBRs will be wrongly
+	 * interpreted. Reset the LBRs here.
+	 *
+	 * Only clear once for a branch counter group with the leader
+	 * event. Because
+	 * - Cannot simply reset the LBRs with the !cpuc->lbr_users.
+	 *   Because it's possible that the last LBR user is not in a
+	 *   branch counter group, e.g., a branch_counters group +
+	 *   several normal LBR events.
+	 * - The LBR reset can be done with any one of the events in a
+	 *   branch counter group, since they are always scheduled together.
+	 *   It's easy to force the leader event an LBR event.
+	 */
+	if (is_branch_counters_group(event) && event == event->group_leader)
+		intel_pmu_lbr_reset();
 }
 
 static inline bool vlbr_exclude_host(void)
···
 	return cycles;
 }
 
+static_assert((64 - PERF_BRANCH_ENTRY_INFO_BITS_MAX) > LBR_INFO_BR_CNTR_NUM * LBR_INFO_BR_CNTR_BITS);
+
 static void intel_pmu_store_lbr(struct cpu_hw_events *cpuc,
 				struct lbr_entry *entries)
 {
···
 		e->abort	= !!(info & LBR_INFO_ABORT);
 		e->cycles	= get_lbr_cycles(info);
 		e->type		= get_lbr_br_type(info);
+
+		/*
+		 * Leverage the reserved field of cpuc->lbr_entries[i] to
+		 * temporarily store the branch counters information.
+		 * The later code will decide what content can be disclosed
+		 * to the perf tool. Pleae see intel_pmu_lbr_counters_reorder().
+		 */
+		e->reserved	= (info >> LBR_INFO_BR_CNTR_OFFSET) & LBR_INFO_BR_CNTR_FULL_MASK;
 	}
 
 	cpuc->lbr_stack.nr = i;
+}
+
+/*
+ * The enabled order may be different from the counter order.
+ * Update the lbr_counters with the enabled order.
+ */
+static void intel_pmu_lbr_counters_reorder(struct cpu_hw_events *cpuc,
+					   struct perf_event *event)
+{
+	int i, j, pos = 0, order[X86_PMC_IDX_MAX];
+	struct perf_event *leader, *sibling;
+	u64 src, dst, cnt;
+
+	leader = event->group_leader;
+	if (branch_sample_counters(leader))
+		order[pos++] = leader->hw.idx;
+
+	for_each_sibling_event(sibling, leader) {
+		if (!branch_sample_counters(sibling))
+			continue;
+		order[pos++] = sibling->hw.idx;
+	}
+
+	WARN_ON_ONCE(!pos);
+
+	for (i = 0; i < cpuc->lbr_stack.nr; i++) {
+		src = cpuc->lbr_entries[i].reserved;
+		dst = 0;
+		for (j = 0; j < pos; j++) {
+			cnt = (src >> (order[j] * LBR_INFO_BR_CNTR_BITS)) & LBR_INFO_BR_CNTR_MASK;
+			dst |= cnt << j * LBR_INFO_BR_CNTR_BITS;
+		}
+		cpuc->lbr_counters[i] = dst;
+		cpuc->lbr_entries[i].reserved = 0;
+	}
+}
+
+void intel_pmu_lbr_save_brstack(struct perf_sample_data *data,
+				struct cpu_hw_events *cpuc,
+				struct perf_event *event)
+{
+	if (is_branch_counters_group(event)) {
+		intel_pmu_lbr_counters_reorder(cpuc, event);
+		perf_sample_save_brstack(data, event, &cpuc->lbr_stack, cpuc->lbr_counters);
+		return;
+	}
+
+	perf_sample_save_brstack(data, event, &cpuc->lbr_stack, NULL);
 }
 
 static void intel_pmu_arch_lbr_read(struct cpu_hw_events *cpuc)
···
 	for (i = 0; i < cpuc->lbr_stack.nr; ) {
 		if (!cpuc->lbr_entries[i].from) {
 			j = i;
-			while (++j < cpuc->lbr_stack.nr)
+			while (++j < cpuc->lbr_stack.nr) {
 				cpuc->lbr_entries[j-1] = cpuc->lbr_entries[j];
+				cpuc->lbr_counters[j-1] = cpuc->lbr_counters[j];
+			}
 			cpuc->lbr_stack.nr--;
 			if (!cpuc->lbr_entries[i].from)
 				continue;
···
 	x86_pmu.lbr_mispred = ecx.split.lbr_mispred;
 	x86_pmu.lbr_timed_lbr = ecx.split.lbr_timed_lbr;
 	x86_pmu.lbr_br_type = ecx.split.lbr_br_type;
+	x86_pmu.lbr_counters = ecx.split.lbr_counters;
 	x86_pmu.lbr_nr = lbr_nr;
+
+	if (!!x86_pmu.lbr_counters)
+		x86_pmu.flags |= PMU_FL_BR_CNTR;
 
 	if (x86_pmu.lbr_mispred)
 		static_branch_enable(&x86_lbr_mispred);

arch/x86/events/intel/uncore.c

···
 	.uncore_units_ignore = spr_uncore_units_ignore,
 };
 
+static const struct intel_uncore_init_fun gnr_uncore_init __initconst = {
+	.cpu_init = gnr_uncore_cpu_init,
+	.pci_init = gnr_uncore_pci_init,
+	.mmio_init = gnr_uncore_mmio_init,
+	.use_discovery = true,
+	.uncore_units_ignore = gnr_uncore_units_ignore,
+};
+
 static const struct intel_uncore_init_fun generic_uncore_init __initconst = {
 	.cpu_init = intel_uncore_generic_uncore_cpu_init,
 	.pci_init = intel_uncore_generic_uncore_pci_init,
···
 	X86_MATCH_INTEL_FAM6_MODEL(METEORLAKE_L,	&mtl_uncore_init),
 	X86_MATCH_INTEL_FAM6_MODEL(SAPPHIRERAPIDS_X,	&spr_uncore_init),
 	X86_MATCH_INTEL_FAM6_MODEL(EMERALDRAPIDS_X,	&spr_uncore_init),
+	X86_MATCH_INTEL_FAM6_MODEL(GRANITERAPIDS_X,	&gnr_uncore_init),
+	X86_MATCH_INTEL_FAM6_MODEL(GRANITERAPIDS_D,	&gnr_uncore_init),
 	X86_MATCH_INTEL_FAM6_MODEL(ATOM_TREMONT_D,	&snr_uncore_init),
 	X86_MATCH_INTEL_FAM6_MODEL(ATOM_GRACEMONT,	&adl_uncore_init),
+	X86_MATCH_INTEL_FAM6_MODEL(ATOM_CRESTMONT_X,	&gnr_uncore_init),
+	X86_MATCH_INTEL_FAM6_MODEL(ATOM_CRESTMONT,	&gnr_uncore_init),
 	{},
 };
 MODULE_DEVICE_TABLE(x86cpu, intel_uncore_match);

arch/x86/events/intel/uncore.h

···
 	unsigned single_fixed:1;
 	unsigned pair_ctr_ctl:1;
 	union {
-		unsigned *msr_offsets;
-		unsigned *pci_offsets;
-		unsigned *mmio_offsets;
+		u64 *msr_offsets;
+		u64 *pci_offsets;
+		u64 *mmio_offsets;
 	};
 	unsigned *box_ids;
 	struct event_constraint unconstrainted;
···
 extern struct pci_extra_dev *uncore_extra_pci_dev;
 extern struct event_constraint uncore_constraint_empty;
 extern int spr_uncore_units_ignore[];
+extern int gnr_uncore_units_ignore[];
 
 /* uncore_snb.c */
 int snb_uncore_pci_init(void);
···
 int spr_uncore_pci_init(void);
 void spr_uncore_cpu_init(void);
 void spr_uncore_mmio_init(void);
+int gnr_uncore_pci_init(void);
+void gnr_uncore_cpu_init(void);
+void gnr_uncore_mmio_init(void);
 
 /* uncore_nhmex.c */
 void nhmex_uncore_cpu_init(void);

arch/x86/events/intel/uncore_discovery.c

···
 		       int die, bool parsed)
 {
 	struct intel_uncore_discovery_type *type;
-	unsigned int *box_offset, *ids;
+	unsigned int *ids;
+	u64 *box_offset;
 	int i;
 
 	if (!unit->ctl || !unit->ctl_offset || !unit->ctr_offset) {
···
 	if (!type)
 		return;
 
-	box_offset = kcalloc(type->num_boxes + 1, sizeof(unsigned int), GFP_KERNEL);
+	box_offset = kcalloc(type->num_boxes + 1, sizeof(u64), GFP_KERNEL);
 	if (!box_offset)
 		return;
 

arch/x86/events/intel/uncore_discovery.h

···
 	u8		ctr_offset;	/* Counter 0 offset */
 	u16		num_boxes;	/* number of boxes for the uncore block */
 	unsigned int	*ids;		/* Box IDs */
-	unsigned int	*box_offset;	/* Box offset */
+	u64		*box_offset;	/* Box offset */
 };
 
 bool intel_uncore_has_discovery_tables(int *ignore);

arch/x86/events/intel/uncore_nhmex.c

···
 };
 
 /* msr offset for each instance of cbox */
-static unsigned nhmex_cbox_msr_offsets[] = {
+static u64 nhmex_cbox_msr_offsets[] = {
 	0x0, 0x80, 0x40, 0xc0, 0x20, 0xa0, 0x60, 0xe0, 0x240, 0x2c0,
 };
 

arch/x86/events/intel/uncore_snbep.c

···
 	return ret;
 }
 
+static int topology_gidnid_map(int nodeid, u32 gidnid)
+{
+	int i, die_id = -1;
+
+	/*
+	 * every three bits in the Node ID mapping register maps
+	 * to a particular node.
+	 */
+	for (i = 0; i < 8; i++) {
+		if (nodeid == GIDNIDMAP(gidnid, i)) {
+			if (topology_max_die_per_package() > 1)
+				die_id = i;
+			else
+				die_id = topology_phys_to_logical_pkg(i);
+			if (die_id < 0)
+				die_id = -ENODEV;
+			break;
+		}
+	}
+
+	return die_id;
+}
+
 /*
  * build pci bus to socket mapping
  */
···
 				break;
 			}
 
-			/*
-			 * every three bits in the Node ID mapping register maps
-			 * to a particular node.
-			 */
-			for (i = 0; i < 8; i++) {
-				if (nodeid == GIDNIDMAP(config, i)) {
-					if (topology_max_die_per_package() > 1)
-						die_id = i;
-					else
-						die_id = topology_phys_to_logical_pkg(i);
-					if (die_id < 0)
-						die_id = -ENODEV;
-					map->pbus_to_dieid[bus] = die_id;
-					break;
-				}
-			}
+			map->pbus_to_dieid[bus] = topology_gidnid_map(nodeid, config);
 			raw_spin_unlock(&pci2phy_map_lock);
 		} else {
 			segment = pci_domain_nr(ubox_dev->bus);
···
 
 /* ICX uncore support */
 
-static unsigned icx_cha_msr_offsets[] = {
+static u64 icx_cha_msr_offsets[] = {
 	0x2a0, 0x2ae, 0x2bc, 0x2ca, 0x2d8, 0x2e6, 0x2f4, 0x302, 0x310,
 	0x31e, 0x32c, 0x33a, 0x348, 0x356, 0x364, 0x372, 0x380, 0x38e,
 	0x3aa, 0x3b8, 0x3c6, 0x3d4, 0x3e2, 0x3f0, 0x3fe, 0x40c, 0x41a,
···
 	.format_group		= &snr_uncore_chabox_format_group,
 };
 
-static unsigned icx_msr_offsets[] = {
+static u64 icx_msr_offsets[] = {
 	0x0, 0x20, 0x40, 0x90, 0xb0, 0xd0,
 };
 
···
 	struct pci_dev *ubox = NULL;
 	struct pci_dev *dev = NULL;
 	u32 nid, gid;
-	int i, idx, ret = -EPERM;
+	int idx, lgc_pkg, ret = -EPERM;
 	struct intel_uncore_topology *upi;
 	unsigned int devfn;
 
···
 			break;
 		}
 
-		for (i = 0; i < 8; i++) {
-			if (nid != GIDNIDMAP(gid, i))
-				continue;
-			for (idx = 0; idx < type->num_boxes; idx++) {
-				upi = &type->topology[nid][idx];
-				devfn = PCI_DEVFN(dev_link0 + idx, ICX_UPI_REGS_ADDR_FUNCTION);
-				dev = pci_get_domain_bus_and_slot(pci_domain_nr(ubox->bus),
-								  ubox->bus->number,
-								  devfn);
-				if (dev) {
-					ret = upi_fill_topology(dev, upi, idx);
-					if (ret)
-						goto err;
-				}
+		lgc_pkg = topology_gidnid_map(nid, gid);
+		if (lgc_pkg < 0) {
+			ret = -EPERM;
+			goto err;
+		}
+		for (idx = 0; idx < type->num_boxes; idx++) {
+			upi = &type->topology[lgc_pkg][idx];
+			devfn = PCI_DEVFN(dev_link0 + idx, ICX_UPI_REGS_ADDR_FUNCTION);
+			dev = pci_get_domain_bus_and_slot(pci_domain_nr(ubox->bus),
+							  ubox->bus->number,
+							  devfn);
+			if (dev) {
+				ret = upi_fill_topology(dev, upi, idx);
+				if (ret)
+					goto err;
 			}
 		}
 	}
···
 	{ /* end: all zeroes */ },
 };
 
+#define SPR_UNCORE_MMIO_COMMON_FORMAT()				\
+	SPR_UNCORE_COMMON_FORMAT(),				\
+	.ops			= &spr_uncore_mmio_ops
+
 static struct intel_uncore_type spr_uncore_imc = {
-	SPR_UNCORE_COMMON_FORMAT(),
+	SPR_UNCORE_MMIO_COMMON_FORMAT(),
 	.name			= "imc",
 	.fixed_ctr_bits		= 48,
 	.fixed_ctr		= SNR_IMC_MMIO_PMON_FIXED_CTR,
 	.fixed_ctl		= SNR_IMC_MMIO_PMON_FIXED_CTL,
-	.ops			= &spr_uncore_mmio_ops,
 	.event_descs		= spr_uncore_imc_events,
 };
 
···
  */
 #define SPR_UNCORE_UPI_NUM_BOXES	4
 
-static unsigned int spr_upi_pci_offsets[SPR_UNCORE_UPI_NUM_BOXES] = {
+static u64 spr_upi_pci_offsets[SPR_UNCORE_UPI_NUM_BOXES] = {
 	0, 0x8000, 0x10000, 0x18000
 };
 
···
 
 static struct intel_uncore_type **
 uncore_get_uncores(enum uncore_access_type type_id, int num_extra,
-		    struct intel_uncore_type **extra)
+		   struct intel_uncore_type **extra, int max_num_types,
+		   struct intel_uncore_type **uncores)
 {
 	struct intel_uncore_type **types, **start_types;
 	int i;
···
 
 	/* Only copy the customized features */
 	for (; *types; types++) {
-		if ((*types)->type_id >= UNCORE_SPR_NUM_UNCORE_TYPES)
+		if ((*types)->type_id >= max_num_types)
 			continue;
-		uncore_type_customized_copy(*types, spr_uncores[(*types)->type_id]);
+		uncore_type_customized_copy(*types, uncores[(*types)->type_id]);
 	}
 
 	for (i = 0; i < num_extra; i++, types++)
···
 
 	uncore_msr_uncores = uncore_get_uncores(UNCORE_ACCESS_MSR,
 						UNCORE_SPR_MSR_EXTRA_UNCORES,
-						spr_msr_uncores);
+						spr_msr_uncores,
+						UNCORE_SPR_NUM_UNCORE_TYPES,
+						spr_uncores);
 
 	type = uncore_find_type_by_id(uncore_msr_uncores, UNCORE_SPR_CHA);
 	if (type) {
···
 	spr_update_device_location(UNCORE_SPR_M3UPI);
 	uncore_pci_uncores = uncore_get_uncores(UNCORE_ACCESS_PCI,
 						UNCORE_SPR_PCI_EXTRA_UNCORES,
-						spr_pci_uncores);
+						spr_pci_uncores,
+						UNCORE_SPR_NUM_UNCORE_TYPES,
+						spr_uncores);
 	return 0;
 }
 
···
 {
 	int ret = snbep_pci2phy_map_init(0x3250, SKX_CPUNODEID, SKX_GIDNIDMAP, true);
 
-	if (ret)
-		uncore_mmio_uncores = uncore_get_uncores(UNCORE_ACCESS_MMIO, 0, NULL);
-	else {
+	if (ret) {
+		uncore_mmio_uncores = uncore_get_uncores(UNCORE_ACCESS_MMIO, 0, NULL,
+							 UNCORE_SPR_NUM_UNCORE_TYPES,
+							 spr_uncores);
+	} else {
 		uncore_mmio_uncores = uncore_get_uncores(UNCORE_ACCESS_MMIO,
 							 UNCORE_SPR_MMIO_EXTRA_UNCORES,
-							 spr_mmio_uncores);
+							 spr_mmio_uncores,
+							 UNCORE_SPR_NUM_UNCORE_TYPES,
+							 spr_uncores);
 
 		spr_uncore_imc_free_running.num_boxes = uncore_type_max_boxes(uncore_mmio_uncores, UNCORE_SPR_IMC) / 2;
 	}
 }
 
 /* end of SPR uncore support */
+
+/* GNR uncore support */
+
+#define UNCORE_GNR_NUM_UNCORE_TYPES	23
+#define UNCORE_GNR_TYPE_15		15
+#define UNCORE_GNR_B2UPI		18
+#define UNCORE_GNR_TYPE_21		21
+#define UNCORE_GNR_TYPE_22		22
+
+int gnr_uncore_units_ignore[] = {
+	UNCORE_SPR_UPI,
+	UNCORE_GNR_TYPE_15,
+	UNCORE_GNR_B2UPI,
+	UNCORE_GNR_TYPE_21,
+	UNCORE_GNR_TYPE_22,
+	UNCORE_IGNORE_END
+};
+
+static struct intel_uncore_type gnr_uncore_ubox = {
+	.name			= "ubox",
+	.attr_update		= uncore_alias_groups,
+};
+
+static struct intel_uncore_type gnr_uncore_b2cmi = {
+	SPR_UNCORE_PCI_COMMON_FORMAT(),
+	.name			= "b2cmi",
+};
+
+static struct intel_uncore_type gnr_uncore_b2cxl = {
+	SPR_UNCORE_MMIO_COMMON_FORMAT(),
+	.name			= "b2cxl",
+};
+
+static struct intel_uncore_type gnr_uncore_mdf_sbo = {
+	.name			= "mdf_sbo",
+	.attr_update		= uncore_alias_groups,
+};
+
+static struct intel_uncore_type *gnr_uncores[UNCORE_GNR_NUM_UNCORE_TYPES] = {
+	&spr_uncore_chabox,
+	&spr_uncore_iio,
+	&spr_uncore_irp,
+	NULL,
+	&spr_uncore_pcu,
+	&gnr_uncore_ubox,
+	&spr_uncore_imc,
+	NULL,
+	NULL,
+	NULL,
+	NULL,
+	NULL,
+	NULL,
+	NULL,
+	NULL,
+	NULL,
+	&gnr_uncore_b2cmi,
+	&gnr_uncore_b2cxl,
+	NULL,
+	NULL,
+	&gnr_uncore_mdf_sbo,
+	NULL,
+	NULL,
+};
+
+static struct freerunning_counters gnr_iio_freerunning[] = {
+	[SPR_IIO_MSR_IOCLK]	= { 0x290e, 0x01, 0x10, 1, 48 },
+	[SPR_IIO_MSR_BW_IN]	= { 0x360e, 0x10, 0x80, 8, 48 },
+	[SPR_IIO_MSR_BW_OUT]	= { 0x2e0e, 0x10, 0x80, 8, 48 },
+};
+
+void gnr_uncore_cpu_init(void)
+{
+	uncore_msr_uncores = uncore_get_uncores(UNCORE_ACCESS_MSR,
+						UNCORE_SPR_MSR_EXTRA_UNCORES,
+						spr_msr_uncores,
+						UNCORE_GNR_NUM_UNCORE_TYPES,
+						gnr_uncores);
+	spr_uncore_iio_free_running.num_boxes = uncore_type_max_boxes(uncore_msr_uncores, UNCORE_SPR_IIO);
+	spr_uncore_iio_free_running.freerunning = gnr_iio_freerunning;
+}
+
+int gnr_uncore_pci_init(void)
+{
+	uncore_pci_uncores = uncore_get_uncores(UNCORE_ACCESS_PCI, 0, NULL,
+						UNCORE_GNR_NUM_UNCORE_TYPES,
+						gnr_uncores);
+	return 0;
+}
+
+void gnr_uncore_mmio_init(void)
+{
+	uncore_mmio_uncores = uncore_get_uncores(UNCORE_ACCESS_MMIO, 0, NULL,
+						 UNCORE_GNR_NUM_UNCORE_TYPES,
+						 gnr_uncores);
+}
+
+/* end of GNR uncore support */

arch/x86/events/perf_event.h

···
 	return is_metric_event(event) || is_slots_event(event);
 }
 
+static inline bool is_branch_counters_group(struct perf_event *event)
+{
+	return event->group_leader->hw.flags & PERF_X86_EVENT_BRANCH_COUNTERS;
+}
+
 struct amd_nb {
 	int nb_id;  /* NorthBridge id */
 	int refcnt; /* reference count */
···
 	int				lbr_pebs_users;
 	struct perf_branch_stack	lbr_stack;
 	struct perf_branch_entry	lbr_entries[MAX_LBR_ENTRIES];
+	u64				lbr_counters[MAX_LBR_ENTRIES]; /* branch stack extra */
 	union {
 		struct er_account		*lbr_sel;
 		struct er_account		*lbr_ctl;
···
 	unsigned int	lbr_mispred:1;
 	unsigned int	lbr_timed_lbr:1;
 	unsigned int	lbr_br_type:1;
+	unsigned int	lbr_counters:4;
 
 	void		(*lbr_reset)(void);
 	void		(*lbr_read)(struct cpu_hw_events *cpuc);
···
 #define PMU_FL_INSTR_LATENCY	0x80 /* Support Instruction Latency in PEBS Memory Info Record */
 #define PMU_FL_MEM_LOADS_AUX	0x100 /* Require an auxiliary event for the complete memory info */
 #define PMU_FL_RETIRE_LATENCY	0x200 /* Support Retire Latency in PEBS */
+#define PMU_FL_BR_CNTR		0x400 /* Support branch counter logging */
 
 #define EVENT_VAR(_id)  event_attr_##_id
 #define EVENT_PTR(_id) &event_attr_##_id.attr.attr
···
 void intel_pmu_store_pebs_lbrs(struct lbr_entry *lbr);
 
 void intel_ds_init(void);
+
+void intel_pmu_lbr_save_brstack(struct perf_sample_data *data,
+				struct cpu_hw_events *cpuc,
+				struct perf_event *event);
 
 void intel_pmu_lbr_swap_task_ctx(struct perf_event_pmu_context *prev_epc,
 				 struct perf_event_pmu_context *next_epc);

arch/x86/events/perf_event_flags.h

···
 PERF_ARCH(PEBS_STLAT,		0x08000) /* st+stlat data address sampling */
 PERF_ARCH(AMD_BRS,		0x10000) /* AMD Branch Sampling */
 PERF_ARCH(PEBS_LAT_HYBRID,	0x20000) /* ld and st lat for hybrid */
+PERF_ARCH(NEEDS_BRANCH_STACK,	0x40000) /* require branch stack setup */
+PERF_ARCH(BRANCH_COUNTERS,	0x80000) /* logs the counters in the extra space of each branch */

arch/x86/include/asm/msr-index.h

···
 #define LBR_INFO_CYCLES			0xffff
 #define LBR_INFO_BR_TYPE_OFFSET		56
 #define LBR_INFO_BR_TYPE		(0xfull << LBR_INFO_BR_TYPE_OFFSET)
+#define LBR_INFO_BR_CNTR_OFFSET		32
+#define LBR_INFO_BR_CNTR_NUM		4
+#define LBR_INFO_BR_CNTR_BITS		2
+#define LBR_INFO_BR_CNTR_MASK		GENMASK_ULL(LBR_INFO_BR_CNTR_BITS - 1, 0)
+#define LBR_INFO_BR_CNTR_FULL_MASK	GENMASK_ULL(LBR_INFO_BR_CNTR_NUM * LBR_INFO_BR_CNTR_BITS - 1, 0)
 
 #define MSR_ARCH_LBR_CTL		0x000014ce
 #define ARCH_LBR_CTL_LBREN		BIT(0)

arch/x86/include/asm/perf_event.h

···
 #define ARCH_PERFMON_EVENTSEL_ENABLE			(1ULL << 22)
 #define ARCH_PERFMON_EVENTSEL_INV			(1ULL << 23)
 #define ARCH_PERFMON_EVENTSEL_CMASK			0xFF000000ULL
+#define ARCH_PERFMON_EVENTSEL_BR_CNTR			(1ULL << 35)
 
 #define INTEL_FIXED_BITS_MASK				0xFULL
 #define INTEL_FIXED_BITS_STRIDE			4
···
 		unsigned int    lbr_timed_lbr:1;
 		/* Branch Type Field Supported */
 		unsigned int    lbr_br_type:1;
+		unsigned int	reserved:13;
+		/* Branch counters (Event Logging) Supported */
+		unsigned int	lbr_counters:4;
 	} split;
 	unsigned int            full;
 };

arch/x86/kernel/smpboot.c

···
 {
 	return cpu_l2c_shared_mask(cpu);
 }
+EXPORT_SYMBOL_GPL(cpu_clustergroup_mask);
 
 static void impress_friends(void)
 {

include/linux/perf_event.h

···
 	return event->attr.branch_sample_type & PERF_SAMPLE_BRANCH_PRIV_SAVE;
 }
 
+static inline bool branch_sample_counters(const struct perf_event *event)
+{
+	return event->attr.branch_sample_type & PERF_SAMPLE_BRANCH_COUNTERS;
+}
+
+static inline bool branch_sample_call_stack(const struct perf_event *event)
+{
+	return event->attr.branch_sample_type & PERF_SAMPLE_BRANCH_CALL_STACK;
+}
 
 struct perf_sample_data {
 	/*
···
 	struct perf_callchain_entry	*callchain;
 	struct perf_raw_record		*raw;
 	struct perf_branch_stack	*br_stack;
+	u64				*br_stack_cntr;
 	union perf_sample_weight	weight;
 	union  perf_mem_data_src	data_src;
 	u64				txn;
···
 
 static inline void perf_sample_save_brstack(struct perf_sample_data *data,
 					    struct perf_event *event,
-					    struct perf_branch_stack *brs)
+					    struct perf_branch_stack *brs,
+					    u64 *brs_cntr)
 {
 	int size = sizeof(u64); /* nr */
 
···
 		size += sizeof(u64);
 	size += brs->nr * sizeof(struct perf_branch_entry);
 
+	/*
+	 * The extension space for counters is appended after the
+	 * struct perf_branch_stack. It is used to store the occurrences
+	 * of events of each branch.
+	 */
+	if (brs_cntr)
+		size += brs->nr * sizeof(u64);
+
 	data->br_stack = brs;
+	data->br_stack_cntr = brs_cntr;
 	data->dyn_size += size;
 	data->sample_flags |= PERF_SAMPLE_BRANCH_STACK;
 }

include/uapi/linux/perf_event.h

···
 
 	PERF_SAMPLE_BRANCH_PRIV_SAVE_SHIFT	= 18, /* save privilege mode */
 
+	PERF_SAMPLE_BRANCH_COUNTERS_SHIFT	= 19, /* save occurrences of events on a branch */
+
 	PERF_SAMPLE_BRANCH_MAX_SHIFT		/* non-ABI */
 };
 
···
 	PERF_SAMPLE_BRANCH_HW_INDEX	= 1U << PERF_SAMPLE_BRANCH_HW_INDEX_SHIFT,
 
 	PERF_SAMPLE_BRANCH_PRIV_SAVE	= 1U << PERF_SAMPLE_BRANCH_PRIV_SAVE_SHIFT,
+
+	PERF_SAMPLE_BRANCH_COUNTERS	= 1U << PERF_SAMPLE_BRANCH_COUNTERS_SHIFT,
 
 	PERF_SAMPLE_BRANCH_MAX		= 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT,
 };
···
 	 *	{ u64                   nr;
 	 *	  { u64	hw_idx; } && PERF_SAMPLE_BRANCH_HW_INDEX
 	 *        { u64 from, to, flags } lbr[nr];
+	 *        #
+	 *        # The format of the counters is decided by the
+	 *        # "branch_counter_nr" and "branch_counter_width",
+	 *        # which are defined in the ABI.
+	 *        #
+	 *        { u64 counters; } cntr[nr] && PERF_SAMPLE_BRANCH_COUNTERS
 	 *      } && PERF_SAMPLE_BRANCH_STACK
 	 *
 	 * 	{ u64			abi; # enum perf_sample_regs_abi
···
 		priv:3,     /* privilege level */
 		reserved:31;
 };
+
+/* Size of used info bits in struct perf_branch_entry */
+#define PERF_BRANCH_ENTRY_INFO_BITS_MAX		33
 
 union perf_sample_weight {
 	__u64		full;

kernel/events/core.c

···
 			if (branch_sample_hw_index(event))
 				perf_output_put(handle, data->br_stack->hw_idx);
 			perf_output_copy(handle, data->br_stack->entries, size);
+			/*
+			 * Add the extension space which is appended
+			 * right after the struct perf_branch_stack.
+			 */
+			if (data->br_stack_cntr) {
+				size = data->br_stack->nr * sizeof(u64);
+				perf_output_copy(handle, data->br_stack_cntr, size);
+			}
 		} else {
 			/*
 			 * we always store at least the value of nr
···
 static struct attribute *pmu_dev_attrs[] = {
 	&dev_attr_type.attr,
 	&dev_attr_perf_event_mux_interval_ms.attr,
+	&dev_attr_nr_addr_filters.attr,
 	NULL,
 };
-ATTRIBUTE_GROUPS(pmu_dev);
+
+static umode_t pmu_dev_is_visible(struct kobject *kobj, struct attribute *a, int n)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct pmu *pmu = dev_get_drvdata(dev);
+
+	if (n == 2 && !pmu->nr_addr_filters)
+		return 0;
+
+	return a->mode;
+}
+
+static struct attribute_group pmu_dev_attr_group = {
+	.is_visible = pmu_dev_is_visible,
+	.attrs = pmu_dev_attrs,
+};
+
+static const struct attribute_group *pmu_dev_groups[] = {
+	&pmu_dev_attr_group,
+	NULL,
+};
 
 static int pmu_bus_running;
 static struct bus_type pmu_bus = {
···
 	if (ret)
 		goto free_dev;
 
-	/* For PMUs with address filters, throw in an extra attribute: */
-	if (pmu->nr_addr_filters)
-		ret = device_create_file(pmu->dev, &dev_attr_nr_addr_filters);
-
-	if (ret)
-		goto del_dev;
-
-	if (pmu->attr_update)
+	if (pmu->attr_update) {
 		ret = sysfs_update_groups(&pmu->dev->kobj, pmu->attr_update);
-
-	if (ret)
-		goto del_dev;
+		if (ret)
+			goto del_dev;
+	}
 
 out:
 	return ret;