Merge tag 'perf-core-2022-08-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull perf events updates from Ingo Molnar:

 - Fix Intel Alder Lake PEBS memory access latency & data source
   profiling info bugs.

 - Use Intel large-PEBS hardware feature in more circumstances, to
   reduce PMI overhead & reduce sampling data.

 - Extend the lost-sample profiling output with the PERF_FORMAT_LOST ABI
   variant, which tells tooling the exact number of samples lost.

 - Add new IBS register bits definitions.

 - AMD uncore events: Add PerfMonV2 DF (Data Fabric) enhancements.

* tag 'perf-core-2022-08-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/ibs: Add new IBS register bits into header
  perf/x86/intel: Fix PEBS data source encoding for ADL
  perf/x86/intel: Fix PEBS memory access info encoding for ADL
  perf/core: Add a new read format to get a number of lost samples
  perf/x86/amd/uncore: Add PerfMonV2 RDPMC assignments
  perf/x86/amd/uncore: Add PerfMonV2 DF event format
  perf/x86/amd/uncore: Detect available DF counters
  perf/x86/amd/uncore: Use attr_update for format attributes
  perf/x86/amd/uncore: Use dynamic events array
  x86/events/intel/ds: Enable large PEBS for PERF_SAMPLE_WEIGHT_TYPE

arch/x86/events/amd/uncore.c

@@ -21,7 +21,6 @@
 #define NUM_COUNTERS_NB		4
 #define NUM_COUNTERS_L2		4
 #define NUM_COUNTERS_L3		6
-#define MAX_COUNTERS		6
 
 #define RDPMC_BASE_NB		6
 #define RDPMC_BASE_LLC		10
@@ -31,6 +30,7 @@
 #undef pr_fmt
 #define pr_fmt(fmt)	"amd_uncore: " fmt
 
+static int pmu_version;
 static int num_counters_llc;
 static int num_counters_nb;
 static bool l3_mask;
@@ -46,7 +46,7 @@ struct amd_uncore {
 	u32 msr_base;
 	cpumask_t *active_mask;
 	struct pmu *pmu;
-	struct perf_event *events[MAX_COUNTERS];
+	struct perf_event **events;
 	struct hlist_node node;
 };
 
@@ -158,6 +158,16 @@ static int amd_uncore_add(struct perf_event *event, int flags)
 	hwc->event_base_rdpmc = uncore->rdpmc_base + hwc->idx;
 	hwc->state = PERF_HES_UPTODATE | PERF_HES_STOPPED;
 
+	/*
+	 * The first four DF counters are accessible via RDPMC index 6 to 9
+	 * followed by the L3 counters from index 10 to 15. For processors
+	 * with more than four DF counters, the DF RDPMC assignments become
+	 * discontiguous as the additional counters are accessible starting
+	 * from index 16.
+	 */
+	if (is_nb_event(event) && hwc->idx >= NUM_COUNTERS_NB)
+		hwc->event_base_rdpmc += NUM_COUNTERS_L3;
+
 	if (flags & PERF_EF_START)
 		amd_uncore_start(event, PERF_EF_RELOAD);
 
@@ -209,10 +219,14 @@ static int amd_uncore_event_init(struct perf_event *event)
 {
 	struct amd_uncore *uncore;
 	struct hw_perf_event *hwc = &event->hw;
+	u64 event_mask = AMD64_RAW_EVENT_MASK_NB;
 
 	if (event->attr.type != event->pmu->type)
 		return -ENOENT;
 
+	if (pmu_version >= 2 && is_nb_event(event))
+		event_mask = AMD64_PERFMON_V2_RAW_EVENT_MASK_NB;
+
 	/*
 	 * NB and Last level cache counters (MSRs) are shared across all cores
 	 * that share the same NB / Last level cache.  On family 16h and below,
@@ -221,7 +235,7 @@ static int amd_uncore_event_init(struct perf_event *event)
 	 * out. So we do not support sampling and per-thread events via
 	 * CAP_NO_INTERRUPT, and we do not enable counter overflow interrupts:
 	 */
-	hwc->config = event->attr.config & AMD64_RAW_EVENT_MASK_NB;
+	hwc->config = event->attr.config & event_mask;
 	hwc->idx = -1;
 
 	if (event->cpu < 0)
@@ -247,6 +261,19 @@ static int amd_uncore_event_init(struct perf_event *event)
 	return 0;
 }
 
+static umode_t
+amd_f17h_uncore_is_visible(struct kobject *kobj, struct attribute *attr, int i)
+{
+	return boot_cpu_data.x86 >= 0x17 && boot_cpu_data.x86 < 0x19 ?
+	       attr->mode : 0;
+}
+
+static umode_t
+amd_f19h_uncore_is_visible(struct kobject *kobj, struct attribute *attr, int i)
+{
+	return boot_cpu_data.x86 >= 0x19 ? attr->mode : 0;
+}
+
 static ssize_t amd_uncore_attr_show_cpumask(struct device *dev,
 					    struct device_attribute *attr,
 					    char *buf)
@@ -287,8 +314,10 @@ static struct device_attribute format_attr_##_var =			\
 
 DEFINE_UNCORE_FORMAT_ATTR(event12,	event,		"config:0-7,32-35");
 DEFINE_UNCORE_FORMAT_ATTR(event14,	event,		"config:0-7,32-35,59-60"); /* F17h+ DF */
+DEFINE_UNCORE_FORMAT_ATTR(event14v2,	event,		"config:0-7,32-37");	   /* PerfMonV2 DF */
 DEFINE_UNCORE_FORMAT_ATTR(event8,	event,		"config:0-7");		   /* F17h+ L3 */
-DEFINE_UNCORE_FORMAT_ATTR(umask,	umask,		"config:8-15");
+DEFINE_UNCORE_FORMAT_ATTR(umask8,	umask,		"config:8-15");
+DEFINE_UNCORE_FORMAT_ATTR(umask12,	umask,		"config:8-15,24-27");	   /* PerfMonV2 DF */
 DEFINE_UNCORE_FORMAT_ATTR(coreid,	coreid,		"config:42-44");	   /* F19h L3 */
 DEFINE_UNCORE_FORMAT_ATTR(slicemask,	slicemask,	"config:48-51");	   /* F17h L3 */
 DEFINE_UNCORE_FORMAT_ATTR(threadmask8,	threadmask,	"config:56-63");	   /* F17h L3 */
@@ -297,20 +326,33 @@ DEFINE_UNCORE_FORMAT_ATTR(enallslices,	enallslices,	"config:46");		   /* F19h L3
 DEFINE_UNCORE_FORMAT_ATTR(enallcores,	enallcores,	"config:47");		   /* F19h L3 */
 DEFINE_UNCORE_FORMAT_ATTR(sliceid,	sliceid,	"config:48-50");	   /* F19h L3 */
 
+/* Common DF and NB attributes */
 static struct attribute *amd_uncore_df_format_attr[] = {
-	&format_attr_event12.attr, /* event14 if F17h+ */
-	&format_attr_umask.attr,
+	&format_attr_event12.attr,	/* event */
+	&format_attr_umask8.attr,	/* umask */
 	NULL,
 };
 
+/* Common L2 and L3 attributes */
 static struct attribute *amd_uncore_l3_format_attr[] = {
-	&format_attr_event12.attr, /* event8 if F17h+ */
-	&format_attr_umask.attr,
-	NULL, /* slicemask if F17h,	coreid if F19h */
-	NULL, /* threadmask8 if F17h,	enallslices if F19h */
-	NULL, /*			enallcores if F19h */
-	NULL, /*			sliceid if F19h */
-	NULL, /*			threadmask2 if F19h */
+	&format_attr_event12.attr,	/* event */
+	&format_attr_umask8.attr,	/* umask */
+	NULL,				/* threadmask */
+	NULL,
+};
+
+/* F17h unique L3 attributes */
+static struct attribute *amd_f17h_uncore_l3_format_attr[] = {
+	&format_attr_slicemask.attr,	/* slicemask */
+	NULL,
+};
+
+/* F19h unique L3 attributes */
+static struct attribute *amd_f19h_uncore_l3_format_attr[] = {
+	&format_attr_coreid.attr,	/* coreid */
+	&format_attr_enallslices.attr,	/* enallslices */
+	&format_attr_enallcores.attr,	/* enallcores */
+	&format_attr_sliceid.attr,	/* sliceid */
 	NULL,
 };
 
@@ -324,6 +366,18 @@ static struct attribute_group amd_uncore_l3_format_group = {
 	.attrs = amd_uncore_l3_format_attr,
 };
 
+static struct attribute_group amd_f17h_uncore_l3_format_group = {
+	.name = "format",
+	.attrs = amd_f17h_uncore_l3_format_attr,
+	.is_visible = amd_f17h_uncore_is_visible,
+};
+
+static struct attribute_group amd_f19h_uncore_l3_format_group = {
+	.name = "format",
+	.attrs = amd_f19h_uncore_l3_format_attr,
+	.is_visible = amd_f19h_uncore_is_visible,
+};
+
 static const struct attribute_group *amd_uncore_df_attr_groups[] = {
 	&amd_uncore_attr_group,
 	&amd_uncore_df_format_group,
@@ -336,6 +390,12 @@ static const struct attribute_group *amd_uncore_l3_attr_groups[] = {
 	NULL,
 };
 
+static const struct attribute_group *amd_uncore_l3_attr_update[] = {
+	&amd_f17h_uncore_l3_format_group,
+	&amd_f19h_uncore_l3_format_group,
+	NULL,
+};
+
 static struct pmu amd_nb_pmu = {
 	.task_ctx_nr	= perf_invalid_context,
 	.attr_groups	= amd_uncore_df_attr_groups,
@@ -353,6 +413,7 @@ static struct pmu amd_nb_pmu = {
 static struct pmu amd_llc_pmu = {
 	.task_ctx_nr	= perf_invalid_context,
 	.attr_groups	= amd_uncore_l3_attr_groups,
+	.attr_update	= amd_uncore_l3_attr_update,
 	.name		= "amd_l2",
 	.event_init	= amd_uncore_event_init,
 	.add		= amd_uncore_add,
@@ -370,11 +431,19 @@ static struct amd_uncore *amd_uncore_alloc(unsigned int cpu)
 			cpu_to_node(cpu));
 }
 
+static inline struct perf_event **
+amd_uncore_events_alloc(unsigned int num, unsigned int cpu)
+{
+	return kzalloc_node(sizeof(struct perf_event *) * num, GFP_KERNEL,
+			    cpu_to_node(cpu));
+}
+
 static int amd_uncore_cpu_up_prepare(unsigned int cpu)
 {
-	struct amd_uncore *uncore_nb = NULL, *uncore_llc;
+	struct amd_uncore *uncore_nb = NULL, *uncore_llc = NULL;
 
 	if (amd_uncore_nb) {
+		*per_cpu_ptr(amd_uncore_nb, cpu) = NULL;
 		uncore_nb = amd_uncore_alloc(cpu);
 		if (!uncore_nb)
 			goto fail;
@@ -384,11 +453,15 @@ 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->events = amd_uncore_events_alloc(num_counters_nb, cpu);
+		if (!uncore_nb->events)
+			goto fail;
 		uncore_nb->id = -1;
 		*per_cpu_ptr(amd_uncore_nb, cpu) = uncore_nb;
 	}
 
 	if (amd_uncore_llc) {
+		*per_cpu_ptr(amd_uncore_llc, cpu) = NULL;
 		uncore_llc = amd_uncore_alloc(cpu);
 		if (!uncore_llc)
 			goto fail;
@@ -398,6 +471,9 @@ static int amd_uncore_cpu_up_prepare(unsigned int cpu)
 		uncore_llc->msr_base = MSR_F16H_L2I_PERF_CTL;
 		uncore_llc->active_mask = &amd_llc_active_mask;
 		uncore_llc->pmu = &amd_llc_pmu;
+		uncore_llc->events = amd_uncore_events_alloc(num_counters_llc, cpu);
+		if (!uncore_llc->events)
+			goto fail;
 		uncore_llc->id = -1;
 		*per_cpu_ptr(amd_uncore_llc, cpu) = uncore_llc;
 	}
@@ -405,9 +481,16 @@ static int amd_uncore_cpu_up_prepare(unsigned int cpu)
 	return 0;
 
 fail:
-	if (amd_uncore_nb)
-		*per_cpu_ptr(amd_uncore_nb, cpu) = NULL;
-	kfree(uncore_nb);
+	if (uncore_nb) {
+		kfree(uncore_nb->events);
+		kfree(uncore_nb);
+	}
+
+	if (uncore_llc) {
+		kfree(uncore_llc->events);
+		kfree(uncore_llc);
+	}
+
 	return -ENOMEM;
 }
 
@@ -540,8 +623,11 @@ static void uncore_dead(unsigned int cpu, struct amd_uncore * __percpu *uncores)
 	if (cpu == uncore->cpu)
 		cpumask_clear_cpu(cpu, uncore->active_mask);
 
-	if (!--uncore->refcnt)
+	if (!--uncore->refcnt) {
+		kfree(uncore->events);
 		kfree(uncore);
+	}
+
 	*per_cpu_ptr(uncores, cpu) = NULL;
 }
 
@@ -560,6 +646,7 @@ static int __init amd_uncore_init(void)
 {
 	struct attribute **df_attr = amd_uncore_df_format_attr;
 	struct attribute **l3_attr = amd_uncore_l3_format_attr;
+	union cpuid_0x80000022_ebx ebx;
 	int ret = -ENODEV;
 
 	if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD &&
@@ -569,6 +656,9 @@ static int __init amd_uncore_init(void)
 	if (!boot_cpu_has(X86_FEATURE_TOPOEXT))
 		return -ENODEV;
 
+	if (boot_cpu_has(X86_FEATURE_PERFMON_V2))
+		pmu_version = 2;
+
 	num_counters_nb	= NUM_COUNTERS_NB;
 	num_counters_llc = NUM_COUNTERS_L2;
 	if (boot_cpu_data.x86 >= 0x17) {
@@ -585,8 +675,12 @@ static int __init amd_uncore_init(void)
 	}
 
 	if (boot_cpu_has(X86_FEATURE_PERFCTR_NB)) {
-		if (boot_cpu_data.x86 >= 0x17)
+		if (pmu_version >= 2) {
+			*df_attr++ = &format_attr_event14v2.attr;
+			*df_attr++ = &format_attr_umask12.attr;
+		} else if (boot_cpu_data.x86 >= 0x17) {
 			*df_attr = &format_attr_event14.attr;
+		}
 
 		amd_uncore_nb = alloc_percpu(struct amd_uncore *);
 		if (!amd_uncore_nb) {
@@ -597,6 +691,11 @@ static int __init amd_uncore_init(void)
 		if (ret)
 			goto fail_nb;
 
+		if (pmu_version >= 2) {
+			ebx.full = cpuid_ebx(EXT_PERFMON_DEBUG_FEATURES);
+			num_counters_nb = ebx.split.num_df_pmc;
+		}
+
 		pr_info("%d %s %s counters detected\n", num_counters_nb,
 			boot_cpu_data.x86_vendor == X86_VENDOR_HYGON ?  "HYGON" : "",
 			amd_nb_pmu.name);
@@ -607,16 +706,11 @@ static int __init amd_uncore_init(void)
 	if (boot_cpu_has(X86_FEATURE_PERFCTR_LLC)) {
 		if (boot_cpu_data.x86 >= 0x19) {
 			*l3_attr++ = &format_attr_event8.attr;
-			*l3_attr++ = &format_attr_umask.attr;
-			*l3_attr++ = &format_attr_coreid.attr;
-			*l3_attr++ = &format_attr_enallslices.attr;
-			*l3_attr++ = &format_attr_enallcores.attr;
-			*l3_attr++ = &format_attr_sliceid.attr;
+			*l3_attr++ = &format_attr_umask8.attr;
 			*l3_attr++ = &format_attr_threadmask2.attr;
 		} else if (boot_cpu_data.x86 >= 0x17) {
 			*l3_attr++ = &format_attr_event8.attr;
-			*l3_attr++ = &format_attr_umask.attr;
-			*l3_attr++ = &format_attr_slicemask.attr;
+			*l3_attr++ = &format_attr_umask8.attr;
 			*l3_attr++ = &format_attr_threadmask8.attr;
 		}
 

arch/x86/events/intel/core.c

@@ -4141,6 +4141,8 @@ tnt_get_event_constraints(struct cpu_hw_events *cpuc, int idx,
 {
 	struct event_constraint *c;
 
+	c = intel_get_event_constraints(cpuc, idx, event);
+
 	/*
 	 * :ppp means to do reduced skid PEBS,
 	 * which is available on PMC0 and fixed counter 0.
@@ -4153,8 +4155,6 @@ tnt_get_event_constraints(struct cpu_hw_events *cpuc, int idx,
 		return &counter0_constraint;
 	}
 
-	c = intel_get_event_constraints(cpuc, idx, event);
-
 	return c;
 }
 
@@ -6241,7 +6241,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.flags |= PMU_FL_INSTR_LATENCY;
 		x86_pmu.flags |= PMU_FL_MEM_LOADS_AUX;
 		x86_pmu.lbr_pt_coexist = true;
-		intel_pmu_pebs_data_source_skl(false);
+		intel_pmu_pebs_data_source_adl();
+		x86_pmu.pebs_latency_data = adl_latency_data_small;
 		x86_pmu.num_topdown_events = 8;
 		x86_pmu.update_topdown_event = adl_update_topdown_event;
 		x86_pmu.set_topdown_event_period = adl_set_topdown_event_period;

arch/x86/events/intel/ds.c

@@ -94,15 +94,40 @@ void __init intel_pmu_pebs_data_source_nhm(void)
 	pebs_data_source[0x07] = OP_LH | P(LVL, L3) | LEVEL(L3) | P(SNOOP, HITM);
 }
 
-void __init intel_pmu_pebs_data_source_skl(bool pmem)
+static void __init __intel_pmu_pebs_data_source_skl(bool pmem, u64 *data_source)
 {
 	u64 pmem_or_l4 = pmem ? LEVEL(PMEM) : LEVEL(L4);
 
-	pebs_data_source[0x08] = OP_LH | pmem_or_l4 | P(SNOOP, HIT);
-	pebs_data_source[0x09] = OP_LH | pmem_or_l4 | REM | P(SNOOP, HIT);
-	pebs_data_source[0x0b] = OP_LH | LEVEL(RAM) | REM | P(SNOOP, NONE);
-	pebs_data_source[0x0c] = OP_LH | LEVEL(ANY_CACHE) | REM | P(SNOOPX, FWD);
-	pebs_data_source[0x0d] = OP_LH | LEVEL(ANY_CACHE) | REM | P(SNOOP, HITM);
+	data_source[0x08] = OP_LH | pmem_or_l4 | P(SNOOP, HIT);
+	data_source[0x09] = OP_LH | pmem_or_l4 | REM | P(SNOOP, HIT);
+	data_source[0x0b] = OP_LH | LEVEL(RAM) | REM | P(SNOOP, NONE);
+	data_source[0x0c] = OP_LH | LEVEL(ANY_CACHE) | REM | P(SNOOPX, FWD);
+	data_source[0x0d] = OP_LH | LEVEL(ANY_CACHE) | REM | P(SNOOP, HITM);
+}
+
+void __init intel_pmu_pebs_data_source_skl(bool pmem)
+{
+	__intel_pmu_pebs_data_source_skl(pmem, pebs_data_source);
+}
+
+static void __init intel_pmu_pebs_data_source_grt(u64 *data_source)
+{
+	data_source[0x05] = OP_LH | P(LVL, L3) | LEVEL(L3) | P(SNOOP, HIT);
+	data_source[0x06] = OP_LH | P(LVL, L3) | LEVEL(L3) | P(SNOOP, HITM);
+	data_source[0x08] = OP_LH | P(LVL, L3) | LEVEL(L3) | P(SNOOPX, FWD);
+}
+
+void __init intel_pmu_pebs_data_source_adl(void)
+{
+	u64 *data_source;
+
+	data_source = x86_pmu.hybrid_pmu[X86_HYBRID_PMU_CORE_IDX].pebs_data_source;
+	memcpy(data_source, pebs_data_source, sizeof(pebs_data_source));
+	__intel_pmu_pebs_data_source_skl(false, data_source);
+
+	data_source = x86_pmu.hybrid_pmu[X86_HYBRID_PMU_ATOM_IDX].pebs_data_source;
+	memcpy(data_source, pebs_data_source, sizeof(pebs_data_source));
+	intel_pmu_pebs_data_source_grt(data_source);
 }
 
 static u64 precise_store_data(u64 status)
@@ -171,7 +196,50 @@ static u64 precise_datala_hsw(struct perf_event *event, u64 status)
 	return dse.val;
 }
 
-static u64 load_latency_data(u64 status)
+static inline void pebs_set_tlb_lock(u64 *val, bool tlb, bool lock)
+{
+	/*
+	 * TLB access
+	 * 0 = did not miss 2nd level TLB
+	 * 1 = missed 2nd level TLB
+	 */
+	if (tlb)
+		*val |= P(TLB, MISS) | P(TLB, L2);
+	else
+		*val |= P(TLB, HIT) | P(TLB, L1) | P(TLB, L2);
+
+	/* locked prefix */
+	if (lock)
+		*val |= P(LOCK, LOCKED);
+}
+
+/* Retrieve the latency data for e-core of ADL */
+u64 adl_latency_data_small(struct perf_event *event, u64 status)
+{
+	union intel_x86_pebs_dse dse;
+	u64 val;
+
+	WARN_ON_ONCE(hybrid_pmu(event->pmu)->cpu_type == hybrid_big);
+
+	dse.val = status;
+
+	val = hybrid_var(event->pmu, pebs_data_source)[dse.ld_dse];
+
+	/*
+	 * For the atom core on ADL,
+	 * bit 4: lock, bit 5: TLB access.
+	 */
+	pebs_set_tlb_lock(&val, dse.ld_locked, dse.ld_stlb_miss);
+
+	if (dse.ld_data_blk)
+		val |= P(BLK, DATA);
+	else
+		val |= P(BLK, NA);
+
+	return val;
+}
+
+static u64 load_latency_data(struct perf_event *event, u64 status)
 {
 	union intel_x86_pebs_dse dse;
 	u64 val;
@@ -181,7 +249,7 @@ static u64 load_latency_data(u64 status)
 	/*
 	 * use the mapping table for bit 0-3
 	 */
-	val = pebs_data_source[dse.ld_dse];
+	val = hybrid_var(event->pmu, pebs_data_source)[dse.ld_dse];
 
 	/*
 	 * Nehalem models do not support TLB, Lock infos
@@ -190,21 +258,8 @@ static u64 load_latency_data(u64 status)
 		val |= P(TLB, NA) | P(LOCK, NA);
 		return val;
 	}
-	/*
-	 * bit 4: TLB access
-	 * 0 = did not miss 2nd level TLB
-	 * 1 = missed 2nd level TLB
-	 */
-	if (dse.ld_stlb_miss)
-		val |= P(TLB, MISS) | P(TLB, L2);
-	else
-		val |= P(TLB, HIT) | P(TLB, L1) | P(TLB, L2);
 
-	/*
-	 * bit 5: locked prefix
-	 */
-	if (dse.ld_locked)
-		val |= P(LOCK, LOCKED);
+	pebs_set_tlb_lock(&val, dse.ld_stlb_miss, dse.ld_locked);
 
 	/*
 	 * Ice Lake and earlier models do not support block infos.
@@ -233,7 +288,7 @@ static u64 load_latency_data(u64 status)
 	return val;
 }
 
-static u64 store_latency_data(u64 status)
+static u64 store_latency_data(struct perf_event *event, u64 status)
 {
 	union intel_x86_pebs_dse dse;
 	u64 val;
@@ -243,23 +298,9 @@ static u64 store_latency_data(u64 status)
 	/*
 	 * use the mapping table for bit 0-3
 	 */
-	val = pebs_data_source[dse.st_lat_dse];
+	val = hybrid_var(event->pmu, pebs_data_source)[dse.st_lat_dse];
 
-	/*
-	 * bit 4: TLB access
-	 * 0 = did not miss 2nd level TLB
-	 * 1 = missed 2nd level TLB
-	 */
-	if (dse.st_lat_stlb_miss)
-		val |= P(TLB, MISS) | P(TLB, L2);
-	else
-		val |= P(TLB, HIT) | P(TLB, L1) | P(TLB, L2);
-
-	/*
-	 * bit 5: locked prefix
-	 */
-	if (dse.st_lat_locked)
-		val |= P(LOCK, LOCKED);
+	pebs_set_tlb_lock(&val, dse.st_lat_stlb_miss, dse.st_lat_locked);
 
 	val |= P(BLK, NA);
 
@@ -781,8 +822,8 @@ struct event_constraint intel_glm_pebs_event_constraints[] = {
 
 struct event_constraint intel_grt_pebs_event_constraints[] = {
 	/* Allow all events as PEBS with no flags */
-	INTEL_PLD_CONSTRAINT(0x5d0, 0xf),
-	INTEL_PSD_CONSTRAINT(0x6d0, 0xf),
+	INTEL_HYBRID_LAT_CONSTRAINT(0x5d0, 0xf),
+	INTEL_HYBRID_LAT_CONSTRAINT(0x6d0, 0xf),
 	EVENT_CONSTRAINT_END
 };
 
@@ -1443,9 +1484,11 @@ static u64 get_data_src(struct perf_event *event, u64 aux)
 	bool fst = fl & (PERF_X86_EVENT_PEBS_ST | PERF_X86_EVENT_PEBS_HSW_PREC);
 
 	if (fl & PERF_X86_EVENT_PEBS_LDLAT)
-		val = load_latency_data(aux);
+		val = load_latency_data(event, aux);
 	else if (fl & PERF_X86_EVENT_PEBS_STLAT)
-		val = store_latency_data(aux);
+		val = store_latency_data(event, aux);
+	else if (fl & PERF_X86_EVENT_PEBS_LAT_HYBRID)
+		val = x86_pmu.pebs_latency_data(event, aux);
 	else if (fst && (fl & PERF_X86_EVENT_PEBS_HSW_PREC))
 		val = precise_datala_hsw(event, aux);
 	else if (fst)

arch/x86/events/perf_event.h

@@ -84,6 +84,7 @@ static inline bool constraint_match(struct event_constraint *c, u64 ecode)
 #define PERF_X86_EVENT_TOPDOWN		0x04000 /* Count Topdown slots/metrics events */
 #define PERF_X86_EVENT_PEBS_STLAT	0x08000 /* st+stlat data address sampling */
 #define PERF_X86_EVENT_AMD_BRS		0x10000 /* AMD Branch Sampling */
+#define PERF_X86_EVENT_PEBS_LAT_HYBRID	0x20000 /* ld and st lat for hybrid */
 
 static inline bool is_topdown_count(struct perf_event *event)
 {
@@ -136,7 +137,8 @@ struct amd_nb {
 	PERF_SAMPLE_DATA_SRC | PERF_SAMPLE_IDENTIFIER | \
 	PERF_SAMPLE_TRANSACTION | PERF_SAMPLE_PHYS_ADDR | \
 	PERF_SAMPLE_REGS_INTR | PERF_SAMPLE_REGS_USER | \
-	PERF_SAMPLE_PERIOD | PERF_SAMPLE_CODE_PAGE_SIZE)
+	PERF_SAMPLE_PERIOD | PERF_SAMPLE_CODE_PAGE_SIZE | \
+	PERF_SAMPLE_WEIGHT_TYPE)
 
 #define PEBS_GP_REGS			\
 	((1ULL << PERF_REG_X86_AX)    | \
@@ -460,6 +462,10 @@ struct cpu_hw_events {
 	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK|X86_ALL_EVENT_FLAGS, \
 			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_ST)
 
+#define INTEL_HYBRID_LAT_CONSTRAINT(c, n)	\
+	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK|X86_ALL_EVENT_FLAGS, \
+			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LAT_HYBRID)
+
 /* Event constraint, but match on all event flags too. */
 #define INTEL_FLAGS_EVENT_CONSTRAINT(c, n) \
 	EVENT_CONSTRAINT(c, n, ARCH_PERFMON_EVENTSEL_EVENT|X86_ALL_EVENT_FLAGS)
@@ -638,6 +644,8 @@ enum {
 	x86_lbr_exclusive_max,
 };
 
+#define PERF_PEBS_DATA_SOURCE_MAX	0x10
+
 struct x86_hybrid_pmu {
 	struct pmu			pmu;
 	const char			*name;
@@ -665,6 +673,8 @@ struct x86_hybrid_pmu {
 	unsigned int			late_ack	:1,
 					mid_ack		:1,
 					enabled_ack	:1;
+
+	u64				pebs_data_source[PERF_PEBS_DATA_SOURCE_MAX];
 };
 
 static __always_inline struct x86_hybrid_pmu *hybrid_pmu(struct pmu *pmu)
@@ -825,6 +835,7 @@ struct x86_pmu {
 	void		(*drain_pebs)(struct pt_regs *regs, struct perf_sample_data *data);
 	struct event_constraint *pebs_constraints;
 	void		(*pebs_aliases)(struct perf_event *event);
+	u64		(*pebs_latency_data)(struct perf_event *event, u64 status);
 	unsigned long	large_pebs_flags;
 	u64		rtm_abort_event;
 
@@ -1392,6 +1403,8 @@ void intel_pmu_disable_bts(void);
 
 int intel_pmu_drain_bts_buffer(void);
 
+u64 adl_latency_data_small(struct perf_event *event, u64 status);
+
 extern struct event_constraint intel_core2_pebs_event_constraints[];
 
 extern struct event_constraint intel_atom_pebs_event_constraints[];
@@ -1499,6 +1512,8 @@ void intel_pmu_pebs_data_source_nhm(void);
 
 void intel_pmu_pebs_data_source_skl(bool pmem);
 
+void intel_pmu_pebs_data_source_adl(void);
+
 int intel_pmu_setup_lbr_filter(struct perf_event *event);
 
 void intel_pt_interrupt(void);

arch/x86/include/asm/amd-ibs.h

@@ -29,7 +29,10 @@ union ibs_fetch_ctl {
 			rand_en:1,	/* 57: random tagging enable */
 			fetch_l2_miss:1,/* 58: L2 miss for sampled fetch
 					 *      (needs IbsFetchComp) */
-			reserved:5;	/* 59-63: reserved */
+			l3_miss_only:1,	/* 59: Collect L3 miss samples only */
+			fetch_oc_miss:1,/* 60: Op cache miss for the sampled fetch */
+			fetch_l3_miss:1,/* 61: L3 cache miss for the sampled fetch */
+			reserved:2;	/* 62-63: reserved */
 	};
 };
 
@@ -38,14 +41,14 @@ union ibs_op_ctl {
 	__u64 val;
 	struct {
 		__u64	opmaxcnt:16,	/* 0-15: periodic op max. count */
-			reserved0:1,	/* 16: reserved */
+			l3_miss_only:1,	/* 16: Collect L3 miss samples only */
 			op_en:1,	/* 17: op sampling enable */
 			op_val:1,	/* 18: op sample valid */
 			cnt_ctl:1,	/* 19: periodic op counter control */
 			opmaxcnt_ext:7,	/* 20-26: upper 7 bits of periodic op maximum count */
-			reserved1:5,	/* 27-31: reserved */
+			reserved0:5,	/* 27-31: reserved */
 			opcurcnt:27,	/* 32-58: periodic op counter current count */
-			reserved2:5;	/* 59-63: reserved */
+			reserved1:5;	/* 59-63: reserved */
 	};
 };
 
@@ -71,11 +74,12 @@ union ibs_op_data {
 union ibs_op_data2 {
 	__u64 val;
 	struct {
-		__u64	data_src:3,	/* 0-2: data source */
+		__u64	data_src_lo:3,	/* 0-2: data source low */
 			reserved0:1,	/* 3: reserved */
 			rmt_node:1,	/* 4: destination node */
 			cache_hit_st:1,	/* 5: cache hit state */
-			reserved1:57;	/* 5-63: reserved */
+			data_src_hi:2,	/* 6-7: data source high */
+			reserved1:56;	/* 8-63: reserved */
 	};
 };
 

arch/x86/include/asm/perf_event.h

@@ -89,6 +89,19 @@
 #define AMD64_RAW_EVENT_MASK_NB		\
 	(AMD64_EVENTSEL_EVENT        |  \
 	 ARCH_PERFMON_EVENTSEL_UMASK)
+
+#define AMD64_PERFMON_V2_EVENTSEL_EVENT_NB	\
+	(AMD64_EVENTSEL_EVENT	|		\
+	 GENMASK_ULL(37, 36))
+
+#define AMD64_PERFMON_V2_EVENTSEL_UMASK_NB	\
+	(ARCH_PERFMON_EVENTSEL_UMASK	|	\
+	 GENMASK_ULL(27, 24))
+
+#define AMD64_PERFMON_V2_RAW_EVENT_MASK_NB		\
+	(AMD64_PERFMON_V2_EVENTSEL_EVENT_NB	|	\
+	 AMD64_PERFMON_V2_EVENTSEL_UMASK_NB)
+
 #define AMD64_NUM_COUNTERS				4
 #define AMD64_NUM_COUNTERS_CORE				6
 #define AMD64_NUM_COUNTERS_NB				4
@@ -194,6 +207,9 @@ union cpuid_0x80000022_ebx {
 	struct {
 		/* Number of Core Performance Counters */
 		unsigned int	num_core_pmc:4;
+		unsigned int	reserved:6;
+		/* Number of Data Fabric Counters */
+		unsigned int	num_df_pmc:6;
 	} split;
 	unsigned int		full;
 };

include/linux/perf_event.h

@@ -759,6 +759,8 @@ struct perf_event {
 	struct pid_namespace		*ns;
 	u64				id;
 
+	atomic64_t			lost_samples;
+
 	u64				(*clock)(void);
 	perf_overflow_handler_t		overflow_handler;
 	void				*overflow_handler_context;

include/uapi/linux/perf_event.h

@@ -301,6 +301,7 @@ enum {
  *	  { u64		time_enabled; } && PERF_FORMAT_TOTAL_TIME_ENABLED
  *	  { u64		time_running; } && PERF_FORMAT_TOTAL_TIME_RUNNING
  *	  { u64		id;           } && PERF_FORMAT_ID
+ *	  { u64		lost;         } && PERF_FORMAT_LOST
  *	} && !PERF_FORMAT_GROUP
  *
  *	{ u64		nr;
@@ -308,6 +309,7 @@ enum {
  *	  { u64		time_running; } && PERF_FORMAT_TOTAL_TIME_RUNNING
  *	  { u64		value;
  *	    { u64	id;           } && PERF_FORMAT_ID
+ *	    { u64	lost;         } && PERF_FORMAT_LOST
  *	  }		cntr[nr];
  *	} && PERF_FORMAT_GROUP
  * };
@@ -317,8 +319,9 @@ enum perf_event_read_format {
 	PERF_FORMAT_TOTAL_TIME_RUNNING		= 1U << 1,
 	PERF_FORMAT_ID				= 1U << 2,
 	PERF_FORMAT_GROUP			= 1U << 3,
+	PERF_FORMAT_LOST			= 1U << 4,
 
-	PERF_FORMAT_MAX = 1U << 4,		/* non-ABI */
+	PERF_FORMAT_MAX = 1U << 5,		/* non-ABI */
 };
 
 #define PERF_ATTR_SIZE_VER0	64	/* sizeof first published struct */

kernel/events/core.c

@@ -1819,6 +1819,9 @@ static void __perf_event_read_size(struct perf_event *event, int nr_siblings)
 	if (event->attr.read_format & PERF_FORMAT_ID)
 		entry += sizeof(u64);
 
+	if (event->attr.read_format & PERF_FORMAT_LOST)
+		entry += sizeof(u64);
+
 	if (event->attr.read_format & PERF_FORMAT_GROUP) {
 		nr += nr_siblings;
 		size += sizeof(u64);
@@ -5260,11 +5263,15 @@ static int __perf_read_group_add(struct perf_event *leader,
 	values[n++] += perf_event_count(leader);
 	if (read_format & PERF_FORMAT_ID)
 		values[n++] = primary_event_id(leader);
+	if (read_format & PERF_FORMAT_LOST)
+		values[n++] = atomic64_read(&leader->lost_samples);
 
 	for_each_sibling_event(sub, leader) {
 		values[n++] += perf_event_count(sub);
 		if (read_format & PERF_FORMAT_ID)
 			values[n++] = primary_event_id(sub);
+		if (read_format & PERF_FORMAT_LOST)
+			values[n++] = atomic64_read(&sub->lost_samples);
 	}
 
 	raw_spin_unlock_irqrestore(&ctx->lock, flags);
@@ -5321,7 +5328,7 @@ static int perf_read_one(struct perf_event *event,
 				 u64 read_format, char __user *buf)
 {
 	u64 enabled, running;
-	u64 values[4];
+	u64 values[5];
 	int n = 0;
 
 	values[n++] = __perf_event_read_value(event, &enabled, &running);
@@ -5331,6 +5338,8 @@ static int perf_read_one(struct perf_event *event,
 		values[n++] = running;
 	if (read_format & PERF_FORMAT_ID)
 		values[n++] = primary_event_id(event);
+	if (read_format & PERF_FORMAT_LOST)
+		values[n++] = atomic64_read(&event->lost_samples);
 
 	if (copy_to_user(buf, values, n * sizeof(u64)))
 		return -EFAULT;
@@ -6858,7 +6867,7 @@ static void perf_output_read_one(struct perf_output_handle *handle,
 				 u64 enabled, u64 running)
 {
 	u64 read_format = event->attr.read_format;
-	u64 values[4];
+	u64 values[5];
 	int n = 0;
 
 	values[n++] = perf_event_count(event);
@@ -6872,6 +6881,8 @@ static void perf_output_read_one(struct perf_output_handle *handle,
 	}
 	if (read_format & PERF_FORMAT_ID)
 		values[n++] = primary_event_id(event);
+	if (read_format & PERF_FORMAT_LOST)
+		values[n++] = atomic64_read(&event->lost_samples);
 
 	__output_copy(handle, values, n * sizeof(u64));
 }
@@ -6882,7 +6893,7 @@ static void perf_output_read_group(struct perf_output_handle *handle,
 {
 	struct perf_event *leader = event->group_leader, *sub;
 	u64 read_format = event->attr.read_format;
-	u64 values[5];
+	u64 values[6];
 	int n = 0;
 
 	values[n++] = 1 + leader->nr_siblings;
@@ -6900,6 +6911,8 @@ static void perf_output_read_group(struct perf_output_handle *handle,
 	values[n++] = perf_event_count(leader);
 	if (read_format & PERF_FORMAT_ID)
 		values[n++] = primary_event_id(leader);
+	if (read_format & PERF_FORMAT_LOST)
+		values[n++] = atomic64_read(&leader->lost_samples);
 
 	__output_copy(handle, values, n * sizeof(u64));
 
@@ -6913,6 +6926,8 @@ static void perf_output_read_group(struct perf_output_handle *handle,
 		values[n++] = perf_event_count(sub);
 		if (read_format & PERF_FORMAT_ID)
 			values[n++] = primary_event_id(sub);
+		if (read_format & PERF_FORMAT_LOST)
+			values[n++] = atomic64_read(&sub->lost_samples);
 
 		__output_copy(handle, values, n * sizeof(u64));
 	}

kernel/events/ring_buffer.c

@@ -172,8 +172,10 @@ __perf_output_begin(struct perf_output_handle *handle,
 		goto out;
 
 	if (unlikely(rb->paused)) {
-		if (rb->nr_pages)
+		if (rb->nr_pages) {
 			local_inc(&rb->lost);
+			atomic64_inc(&event->lost_samples);
+		}
 		goto out;
 	}
 
@@ -254,6 +256,7 @@ __perf_output_begin(struct perf_output_handle *handle,
 
 fail:
 	local_inc(&rb->lost);
+	atomic64_inc(&event->lost_samples);
 	perf_output_put_handle(handle);
 out:
 	rcu_read_unlock();