Merge tag 'perf-core-2022-12-12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull perf events updates from Ingo Molnar:

 - Thoroughly rewrite the data structures that implement perf task
   context handling, with the goal of fixing various quirks and
   unfeatures both in already merged, and in upcoming proposed code.

   The old data structure is the per task and per cpu
   perf_event_contexts:

         task_struct::perf_events_ctxp[] <-> perf_event_context <-> perf_cpu_context
              ^                                 |    ^     |           ^
              `---------------------------------'    |     `--> pmu ---'
                                                     v           ^
                                                perf_event ------'

   In this new design this is replaced with a single task context and a
   single CPU context, plus intermediate data-structures:

         task_struct::perf_event_ctxp -> perf_event_context <- perf_cpu_context
              ^                           |   ^ ^
              `---------------------------'   | |
                                              | |    perf_cpu_pmu_context <--.
                                              | `----.    ^                  |
                                              |      |    |                  |
                                              |      v    v                  |
                                              | ,--> perf_event_pmu_context  |
                                              | |                            |
                                              | |                            |
                                              v v                            |
                                         perf_event ---> pmu ----------------'

   [ See commit bd2756811766 for more details. ]

   This rewrite was developed by Peter Zijlstra and Ravi Bangoria.

 - Optimize perf_tp_event()

 - Update the Intel uncore PMU driver, extending it with UPI topology
   discovery on various hardware models.

 - Misc fixes & cleanups

* tag 'perf-core-2022-12-12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (25 commits)
  perf/x86/intel/uncore: Fix reference count leak in __uncore_imc_init_box()
  perf/x86/intel/uncore: Fix reference count leak in snr_uncore_mmio_map()
  perf/x86/intel/uncore: Fix reference count leak in hswep_has_limit_sbox()
  perf/x86/intel/uncore: Fix reference count leak in sad_cfg_iio_topology()
  perf/x86/intel/uncore: Make set_mapping() procedure void
  perf/x86/intel/uncore: Update sysfs-devices-mapping file
  perf/x86/intel/uncore: Enable UPI topology discovery for Sapphire Rapids
  perf/x86/intel/uncore: Enable UPI topology discovery for Icelake Server
  perf/x86/intel/uncore: Get UPI NodeID and GroupID
  perf/x86/intel/uncore: Enable UPI topology discovery for Skylake Server
  perf/x86/intel/uncore: Generalize get_topology() for SKX PMUs
  perf/x86/intel/uncore: Disable I/O stacks to PMU mapping on ICX-D
  perf/x86/intel/uncore: Clear attr_update properly
  perf/x86/intel/uncore: Introduce UPI topology type
  perf/x86/intel/uncore: Generalize IIO topology support
  perf/core: Don't allow grouping events from different hw pmus
  perf/amd/ibs: Make IBS a core pmu
  perf: Fix function pointer case
  perf/x86/amd: Remove the repeated declaration
  perf: Fix possible memleak in pmu_dev_alloc()
  ...

6 files changed, 485 insertions, 94 deletions

diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c
index 1b92bf05fd65..dfd2c124cdf8 100644
--- a/arch/x86/events/intel/core.c
+++ b/arch/x86/events/intel/core.c
@@ -4536,8 +4536,6 @@ end:
 	cpumask_set_cpu(cpu, &pmu->supported_cpus);
 	cpuc->pmu = &pmu->pmu;
 
-	x86_pmu_update_cpu_context(&pmu->pmu, cpu);
-
 	return true;
 }
 
@@ -4671,17 +4669,17 @@ static void intel_pmu_cpu_dead(int cpu)
 		cpumask_clear_cpu(cpu, &hybrid_pmu(cpuc->pmu)->supported_cpus);
 }
 
-static void intel_pmu_sched_task(struct perf_event_context *ctx,
+static void intel_pmu_sched_task(struct perf_event_pmu_context *pmu_ctx,
 				 bool sched_in)
 {
-	intel_pmu_pebs_sched_task(ctx, sched_in);
-	intel_pmu_lbr_sched_task(ctx, sched_in);
+	intel_pmu_pebs_sched_task(pmu_ctx, sched_in);
+	intel_pmu_lbr_sched_task(pmu_ctx, sched_in);
 }
 
-static void intel_pmu_swap_task_ctx(struct perf_event_context *prev,
-				    struct perf_event_context *next)
+static void intel_pmu_swap_task_ctx(struct perf_event_pmu_context *prev_epc,
+				    struct perf_event_pmu_context *next_epc)
 {
-	intel_pmu_lbr_swap_task_ctx(prev, next);
+	intel_pmu_lbr_swap_task_ctx(prev_epc, next_epc);
 }
 
 static int intel_pmu_check_period(struct perf_event *event, u64 value)
@@ -4705,12 +4703,11 @@ static int intel_pmu_aux_output_match(struct perf_event *event)
 	return is_intel_pt_event(event);
 }
 
-static int intel_pmu_filter_match(struct perf_event *event)
+static void intel_pmu_filter(struct pmu *pmu, int cpu, bool *ret)
 {
-	struct x86_hybrid_pmu *pmu = hybrid_pmu(event->pmu);
-	unsigned int cpu = smp_processor_id();
+	struct x86_hybrid_pmu *hpmu = hybrid_pmu(pmu);
 
-	return cpumask_test_cpu(cpu, &pmu->supported_cpus);
+	*ret = !cpumask_test_cpu(cpu, &hpmu->supported_cpus);
 }
 
 PMU_FORMAT_ATTR(offcore_rsp, "config1:0-63");
@@ -6413,7 +6410,7 @@ __init int intel_pmu_init(void)
 		static_call_update(intel_pmu_set_topdown_event_period,
 				   &adl_set_topdown_event_period);
 
-		x86_pmu.filter_match = intel_pmu_filter_match;
+		x86_pmu.filter = intel_pmu_filter;
 		x86_pmu.get_event_constraints = adl_get_event_constraints;
 		x86_pmu.hw_config = adl_hw_config;
 		x86_pmu.limit_period = spr_limit_period;
diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c
index 446d2833efa7..88e58b6ee73c 100644
--- a/arch/x86/events/intel/ds.c
+++ b/arch/x86/events/intel/ds.c
@@ -1069,7 +1069,7 @@ static inline bool pebs_needs_sched_cb(struct cpu_hw_events *cpuc)
 	return cpuc->n_pebs && (cpuc->n_pebs == cpuc->n_large_pebs);
 }
 
-void intel_pmu_pebs_sched_task(struct perf_event_context *ctx, bool sched_in)
+void intel_pmu_pebs_sched_task(struct perf_event_pmu_context *pmu_ctx, bool sched_in)
 {
 	struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events);
 
@@ -1177,7 +1177,7 @@ static void
 pebs_update_state(bool needed_cb, struct cpu_hw_events *cpuc,
 		  struct perf_event *event, bool add)
 {
-	struct pmu *pmu = event->ctx->pmu;
+	struct pmu *pmu = event->pmu;
 	/*
 	 * Make sure we get updated with the first PEBS
 	 * event. It will trigger also during removal, but
diff --git a/arch/x86/events/intel/lbr.c b/arch/x86/events/intel/lbr.c
index 8259d725054d..017baba56b01 100644
--- a/arch/x86/events/intel/lbr.c
+++ b/arch/x86/events/intel/lbr.c
@@ -515,21 +515,21 @@ static void __intel_pmu_lbr_save(void *ctx)
 	cpuc->last_log_id = ++task_context_opt(ctx)->log_id;
 }
 
-void intel_pmu_lbr_swap_task_ctx(struct perf_event_context *prev,
-				 struct perf_event_context *next)
+void intel_pmu_lbr_swap_task_ctx(struct perf_event_pmu_context *prev_epc,
+				 struct perf_event_pmu_context *next_epc)
 {
 	void *prev_ctx_data, *next_ctx_data;
 
-	swap(prev->task_ctx_data, next->task_ctx_data);
+	swap(prev_epc->task_ctx_data, next_epc->task_ctx_data);
 
 	/*
-	 * Architecture specific synchronization makes sense in
-	 * case both prev->task_ctx_data and next->task_ctx_data
+	 * Architecture specific synchronization makes sense in case
+	 * both prev_epc->task_ctx_data and next_epc->task_ctx_data
 	 * pointers are allocated.
 	 */
 
-	prev_ctx_data = next->task_ctx_data;
-	next_ctx_data = prev->task_ctx_data;
+	prev_ctx_data = next_epc->task_ctx_data;
+	next_ctx_data = prev_epc->task_ctx_data;
 
 	if (!prev_ctx_data || !next_ctx_data)
 		return;
@@ -538,7 +538,7 @@ void intel_pmu_lbr_swap_task_ctx(struct perf_event_context *prev,
 	     task_context_opt(next_ctx_data)->lbr_callstack_users);
 }
 
-void intel_pmu_lbr_sched_task(struct perf_event_context *ctx, bool sched_in)
+void intel_pmu_lbr_sched_task(struct perf_event_pmu_context *pmu_ctx, bool sched_in)
 {
 	struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events);
 	void *task_ctx;
@@ -551,7 +551,7 @@ void intel_pmu_lbr_sched_task(struct perf_event_context *ctx, bool sched_in)
 	 * the task was scheduled out, restore the stack. Otherwise flush
 	 * the LBR stack.
 	 */
-	task_ctx = ctx ? ctx->task_ctx_data : NULL;
+	task_ctx = pmu_ctx ? pmu_ctx->task_ctx_data : NULL;
 	if (task_ctx) {
 		if (sched_in)
 			__intel_pmu_lbr_restore(task_ctx);
@@ -587,8 +587,8 @@ void intel_pmu_lbr_add(struct perf_event *event)
 
 	cpuc->br_sel = event->hw.branch_reg.reg;
 
-	if (branch_user_callstack(cpuc->br_sel) && event->ctx->task_ctx_data)
-		task_context_opt(event->ctx->task_ctx_data)->lbr_callstack_users++;
+	if (branch_user_callstack(cpuc->br_sel) && event->pmu_ctx->task_ctx_data)
+		task_context_opt(event->pmu_ctx->task_ctx_data)->lbr_callstack_users++;
 
 	/*
 	 * Request pmu::sched_task() callback, which will fire inside the
@@ -611,7 +611,7 @@ void intel_pmu_lbr_add(struct perf_event *event)
 	 */
 	if (x86_pmu.intel_cap.pebs_baseline && event->attr.precise_ip > 0)
 		cpuc->lbr_pebs_users++;
-	perf_sched_cb_inc(event->ctx->pmu);
+	perf_sched_cb_inc(event->pmu);
 	if (!cpuc->lbr_users++ && !event->total_time_running)
 		intel_pmu_lbr_reset();
 }
@@ -664,8 +664,8 @@ void intel_pmu_lbr_del(struct perf_event *event)
 		return;
 
 	if (branch_user_callstack(cpuc->br_sel) &&
-	    event->ctx->task_ctx_data)
-		task_context_opt(event->ctx->task_ctx_data)->lbr_callstack_users--;
+	    event->pmu_ctx->task_ctx_data)
+		task_context_opt(event->pmu_ctx->task_ctx_data)->lbr_callstack_users--;
 
 	if (event->hw.flags & PERF_X86_EVENT_LBR_SELECT)
 		cpuc->lbr_select = 0;
@@ -675,7 +675,7 @@ void intel_pmu_lbr_del(struct perf_event *event)
 	cpuc->lbr_users--;
 	WARN_ON_ONCE(cpuc->lbr_users < 0);
 	WARN_ON_ONCE(cpuc->lbr_pebs_users < 0);
-	perf_sched_cb_dec(event->ctx->pmu);
+	perf_sched_cb_dec(event->pmu);
 }
 
 static inline bool vlbr_exclude_host(void)
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 2adeaf4de4df..e278e2e7c051 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -2,6 +2,7 @@
 #include <linux/slab.h>
 #include <linux/pci.h>
 #include <asm/apicdef.h>
+#include <asm/intel-family.h>
 #include <linux/io-64-nonatomic-lo-hi.h>
 
 #include <linux/perf_event.h>
@@ -88,12 +89,12 @@ struct intel_uncore_type {
 	 * to identify which platform component each PMON block of that type is
 	 * supposed to monitor.
 	 */
-	struct intel_uncore_topology *topology;
+	struct intel_uncore_topology **topology;
 	/*
 	 * Optional callbacks for managing mapping of Uncore units to PMONs
 	 */
 	int (*get_topology)(struct intel_uncore_type *type);
-	int (*set_mapping)(struct intel_uncore_type *type);
+	void (*set_mapping)(struct intel_uncore_type *type);
 	void (*cleanup_mapping)(struct intel_uncore_type *type);
 };
 
@@ -178,11 +179,26 @@ struct freerunning_counters {
 	unsigned *box_offsets;
 };
 
-struct intel_uncore_topology {
-	u64 configuration;
+struct uncore_iio_topology {
+	int pci_bus_no;
 	int segment;
 };
 
+struct uncore_upi_topology {
+	int die_to;
+	int pmu_idx_to;
+	int enabled;
+};
+
+struct intel_uncore_topology {
+	int pmu_idx;
+	union {
+		void *untyped;
+		struct uncore_iio_topology *iio;
+		struct uncore_upi_topology *upi;
+	};
+};
+
 struct pci2phy_map {
 	struct list_head list;
 	int segment;
diff --git a/arch/x86/events/intel/uncore_snb.c b/arch/x86/events/intel/uncore_snb.c
index 1ef4f7861e2e..1f4869227efb 100644
--- a/arch/x86/events/intel/uncore_snb.c
+++ b/arch/x86/events/intel/uncore_snb.c
@@ -1338,6 +1338,7 @@ static void __uncore_imc_init_box(struct intel_uncore_box *box,
 	/* MCHBAR is disabled */
 	if (!(mch_bar & BIT(0))) {
 		pr_warn("perf uncore: MCHBAR is disabled. Failed to map IMC free-running counters.\n");
+		pci_dev_put(pdev);
 		return;
 	}
 	mch_bar &= ~BIT(0);
@@ -1352,6 +1353,8 @@ static void __uncore_imc_init_box(struct intel_uncore_box *box,
 	box->io_addr = ioremap(addr, type->mmio_map_size);
 	if (!box->io_addr)
 		pr_warn("perf uncore: Failed to ioremap for %s.\n", type->name);
+
+	pci_dev_put(pdev);
 }
 
 static void tgl_uncore_imc_freerunning_init_box(struct intel_uncore_box *box)
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index ed869443efb2..44c2f879f708 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -445,6 +445,7 @@
 #define ICX_UPI_PCI_PMON_CTR0			0x320
 #define ICX_UPI_PCI_PMON_BOX_CTL		0x318
 #define ICX_UPI_CTL_UMASK_EXT			0xffffff
+#define ICX_UBOX_DID				0x3450
 
 /* ICX M3UPI*/
 #define ICX_M3UPI_PCI_PMON_CTL0			0xd8
@@ -457,6 +458,7 @@
 
 /* SPR */
 #define SPR_RAW_EVENT_MASK_EXT			0xffffff
+#define SPR_UBOX_DID				0x3250
 
 /* SPR CHA */
 #define SPR_CHA_PMON_CTL_TID_EN			(1 << 16)
@@ -1372,6 +1374,28 @@ static struct pci_driver snbep_uncore_pci_driver = {
 
 #define NODE_ID_MASK	0x7
 
+/* Each three bits from 0 to 23 of GIDNIDMAP register correspond Node ID. */
+#define GIDNIDMAP(config, id)	(((config) >> (3 * (id))) & 0x7)
+
+static int upi_nodeid_groupid(struct pci_dev *ubox_dev, int nodeid_loc, int idmap_loc,
+			      int *nodeid, int *groupid)
+{
+	int ret;
+
+	/* get the Node ID of the local register */
+	ret = pci_read_config_dword(ubox_dev, nodeid_loc, nodeid);
+	if (ret)
+		goto err;
+
+	*nodeid = *nodeid & NODE_ID_MASK;
+	/* get the Node ID mapping */
+	ret = pci_read_config_dword(ubox_dev, idmap_loc, groupid);
+	if (ret)
+		goto err;
+err:
+	return ret;
+}
+
 /*
  * build pci bus to socket mapping
  */
@@ -1397,13 +1421,8 @@ static int snbep_pci2phy_map_init(int devid, int nodeid_loc, int idmap_loc, bool
 		 * the topology.
 		 */
 		if (nr_node_ids <= 8) {
-			/* get the Node ID of the local register */
-			err = pci_read_config_dword(ubox_dev, nodeid_loc, &config);
-			if (err)
-				break;
-			nodeid = config & NODE_ID_MASK;
-			/* get the Node ID mapping */
-			err = pci_read_config_dword(ubox_dev, idmap_loc, &config);
+			err = upi_nodeid_groupid(ubox_dev, nodeid_loc, idmap_loc,
+						 &nodeid, &config);
 			if (err)
 				break;
 
@@ -1421,7 +1440,7 @@ static int snbep_pci2phy_map_init(int devid, int nodeid_loc, int idmap_loc, bool
 			 * to a particular node.
 			 */
 			for (i = 0; i < 8; i++) {
-				if (nodeid == ((config >> (3 * i)) & 0x7)) {
+				if (nodeid == GIDNIDMAP(config, i)) {
 					if (topology_max_die_per_package() > 1)
 						die_id = i;
 					else
@@ -2891,6 +2910,7 @@ static bool hswep_has_limit_sbox(unsigned int device)
 		return false;
 
 	pci_read_config_dword(dev, HSWEP_PCU_CAPID4_OFFET, &capid4);
+	pci_dev_put(dev);
 	if (!hswep_get_chop(capid4))
 		return true;
 
@@ -3699,10 +3719,16 @@ static struct intel_uncore_ops skx_uncore_iio_ops = {
 	.read_counter		= uncore_msr_read_counter,
 };
 
-static inline u8 skx_iio_stack(struct intel_uncore_pmu *pmu, int die)
+static struct intel_uncore_topology *pmu_topology(struct intel_uncore_pmu *pmu, int die)
 {
-	return pmu->type->topology[die].configuration >>
-	       (pmu->pmu_idx * BUS_NUM_STRIDE);
+	int idx;
+
+	for (idx = 0; idx < pmu->type->num_boxes; idx++) {
+		if (pmu->type->topology[die][idx].pmu_idx == pmu->pmu_idx)
+			return &pmu->type->topology[die][idx];
+	}
+
+	return NULL;
 }
 
 static umode_t
@@ -3710,8 +3736,9 @@ pmu_iio_mapping_visible(struct kobject *kobj, struct attribute *attr,
 			 int die, int zero_bus_pmu)
 {
 	struct intel_uncore_pmu *pmu = dev_to_uncore_pmu(kobj_to_dev(kobj));
+	struct intel_uncore_topology *pmut = pmu_topology(pmu, die);
 
-	return (!skx_iio_stack(pmu, die) && pmu->pmu_idx != zero_bus_pmu) ? 0 : attr->mode;
+	return (pmut && !pmut->iio->pci_bus_no && pmu->pmu_idx != zero_bus_pmu) ? 0 : attr->mode;
 }
 
 static umode_t
@@ -3727,9 +3754,10 @@ static ssize_t skx_iio_mapping_show(struct device *dev,
 	struct intel_uncore_pmu *pmu = dev_to_uncore_pmu(dev);
 	struct dev_ext_attribute *ea = to_dev_ext_attribute(attr);
 	long die = (long)ea->var;
+	struct intel_uncore_topology *pmut = pmu_topology(pmu, die);
 
-	return sprintf(buf, "%04x:%02x\n", pmu->type->topology[die].segment,
-					   skx_iio_stack(pmu, die));
+	return sprintf(buf, "%04x:%02x\n", pmut ? pmut->iio->segment : 0,
+					   pmut ? pmut->iio->pci_bus_no : 0);
 }
 
 static int skx_msr_cpu_bus_read(int cpu, u64 *topology)
@@ -3764,18 +3792,79 @@ static int die_to_cpu(int die)
 	return res;
 }
 
-static int skx_iio_get_topology(struct intel_uncore_type *type)
+enum {
+	IIO_TOPOLOGY_TYPE,
+	UPI_TOPOLOGY_TYPE,
+	TOPOLOGY_MAX
+};
+
+static const size_t topology_size[TOPOLOGY_MAX] = {
+	sizeof(*((struct intel_uncore_topology *)NULL)->iio),
+	sizeof(*((struct intel_uncore_topology *)NULL)->upi)
+};
+
+static int pmu_alloc_topology(struct intel_uncore_type *type, int topology_type)
 {
-	int die, ret = -EPERM;
+	int die, idx;
+	struct intel_uncore_topology **topology;
+
+	if (!type->num_boxes)
+		return -EPERM;
 
-	type->topology = kcalloc(uncore_max_dies(), sizeof(*type->topology),
-				 GFP_KERNEL);
-	if (!type->topology)
-		return -ENOMEM;
+	topology = kcalloc(uncore_max_dies(), sizeof(*topology), GFP_KERNEL);
+	if (!topology)
+		goto err;
 
 	for (die = 0; die < uncore_max_dies(); die++) {
-		ret = skx_msr_cpu_bus_read(die_to_cpu(die),
-					   &type->topology[die].configuration);
+		topology[die] = kcalloc(type->num_boxes, sizeof(**topology), GFP_KERNEL);
+		if (!topology[die])
+			goto clear;
+		for (idx = 0; idx < type->num_boxes; idx++) {
+			topology[die][idx].untyped = kcalloc(type->num_boxes,
+							     topology_size[topology_type],
+							     GFP_KERNEL);
+			if (!topology[die][idx].untyped)
+				goto clear;
+		}
+	}
+
+	type->topology = topology;
+
+	return 0;
+clear:
+	for (; die >= 0; die--) {
+		for (idx = 0; idx < type->num_boxes; idx++)
+			kfree(topology[die][idx].untyped);
+		kfree(topology[die]);
+	}
+	kfree(topology);
+err:
+	return -ENOMEM;
+}
+
+static void pmu_free_topology(struct intel_uncore_type *type)
+{
+	int die, idx;
+
+	if (type->topology) {
+		for (die = 0; die < uncore_max_dies(); die++) {
+			for (idx = 0; idx < type->num_boxes; idx++)
+				kfree(type->topology[die][idx].untyped);
+			kfree(type->topology[die]);
+		}
+		kfree(type->topology);
+		type->topology = NULL;
+	}
+}
+
+static int skx_pmu_get_topology(struct intel_uncore_type *type,
+				 int (*topology_cb)(struct intel_uncore_type*, int, int, u64))
+{
+	int die, ret = -EPERM;
+	u64 cpu_bus_msr;
+
+	for (die = 0; die < uncore_max_dies(); die++) {
+		ret = skx_msr_cpu_bus_read(die_to_cpu(die), &cpu_bus_msr);
 		if (ret)
 			break;
 
@@ -3783,15 +3872,33 @@ static int skx_iio_get_topology(struct intel_uncore_type *type)
 		if (ret < 0)
 			break;
 
-		type->topology[die].segment = ret;
+		ret = topology_cb(type, ret, die, cpu_bus_msr);
+		if (ret)
+			break;
 	}
 
-	if (ret < 0) {
-		kfree(type->topology);
-		type->topology = NULL;
+	return ret;
+}
+
+static int skx_iio_topology_cb(struct intel_uncore_type *type, int segment,
+				int die, u64 cpu_bus_msr)
+{
+	int idx;
+	struct intel_uncore_topology *t;
+
+	for (idx = 0; idx < type->num_boxes; idx++) {
+		t = &type->topology[die][idx];
+		t->pmu_idx = idx;
+		t->iio->segment = segment;
+		t->iio->pci_bus_no = (cpu_bus_msr >> (idx * BUS_NUM_STRIDE)) & 0xff;
 	}
 
-	return ret;
+	return 0;
+}
+
+static int skx_iio_get_topology(struct intel_uncore_type *type)
+{
+	return skx_pmu_get_topology(type, skx_iio_topology_cb);
 }
 
 static struct attribute_group skx_iio_mapping_group = {
@@ -3803,8 +3910,25 @@ static const struct attribute_group *skx_iio_attr_update[] = {
 	NULL,
 };
 
-static int
-pmu_iio_set_mapping(struct intel_uncore_type *type, struct attribute_group *ag)
+static void pmu_clear_mapping_attr(const struct attribute_group **groups,
+				   struct attribute_group *ag)
+{
+	int i;
+
+	for (i = 0; groups[i]; i++) {
+		if (groups[i] == ag) {
+			for (i++; groups[i]; i++)
+				groups[i - 1] = groups[i];
+			groups[i - 1] = NULL;
+			break;
+		}
+	}
+}
+
+static void
+pmu_set_mapping(struct intel_uncore_type *type, struct attribute_group *ag,
+		ssize_t (*show)(struct device*, struct device_attribute*, char*),
+		int topology_type)
 {
 	char buf[64];
 	int ret;
@@ -3812,11 +3936,13 @@ pmu_iio_set_mapping(struct intel_uncore_type *type, struct attribute_group *ag)
 	struct attribute **attrs = NULL;
 	struct dev_ext_attribute *eas = NULL;
 
-	ret = type->get_topology(type);
+	ret = pmu_alloc_topology(type, topology_type);
 	if (ret < 0)
 		goto clear_attr_update;
 
-	ret = -ENOMEM;
+	ret = type->get_topology(type);
+	if (ret < 0)
+		goto clear_topology;
 
 	/* One more for NULL. */
 	attrs = kcalloc((uncore_max_dies() + 1), sizeof(*attrs), GFP_KERNEL);
@@ -3828,20 +3954,20 @@ pmu_iio_set_mapping(struct intel_uncore_type *type, struct attribute_group *ag)
 		goto clear_attrs;
 
 	for (die = 0; die < uncore_max_dies(); die++) {
-		sprintf(buf, "die%ld", die);
+		snprintf(buf, sizeof(buf), "die%ld", die);
 		sysfs_attr_init(&eas[die].attr.attr);
 		eas[die].attr.attr.name = kstrdup(buf, GFP_KERNEL);
 		if (!eas[die].attr.attr.name)
 			goto err;
 		eas[die].attr.attr.mode = 0444;
-		eas[die].attr.show = skx_iio_mapping_show;
+		eas[die].attr.show = show;
 		eas[die].attr.store = NULL;
 		eas[die].var = (void *)die;
 		attrs[die] = &eas[die].attr.attr;
 	}
 	ag->attrs = attrs;
 
-	return 0;
+	return;
 err:
 	for (; die >= 0; die--)
 		kfree(eas[die].attr.attr.name);
@@ -3849,14 +3975,13 @@ err:
 clear_attrs:
 	kfree(attrs);
 clear_topology:
-	kfree(type->topology);
+	pmu_free_topology(type);
 clear_attr_update:
-	type->attr_update = NULL;
-	return ret;
+	pmu_clear_mapping_attr(type->attr_update, ag);
 }
 
 static void
-pmu_iio_cleanup_mapping(struct intel_uncore_type *type, struct attribute_group *ag)
+pmu_cleanup_mapping(struct intel_uncore_type *type, struct attribute_group *ag)
 {
 	struct attribute **attr = ag->attrs;
 
@@ -3868,17 +3993,23 @@ pmu_iio_cleanup_mapping(struct intel_uncore_type *type, struct attribute_group *
 	kfree(attr_to_ext_attr(*ag->attrs));
 	kfree(ag->attrs);
 	ag->attrs = NULL;
-	kfree(type->topology);
+	pmu_free_topology(type);
+}
+
+static void
+pmu_iio_set_mapping(struct intel_uncore_type *type, struct attribute_group *ag)
+{
+	pmu_set_mapping(type, ag, skx_iio_mapping_show, IIO_TOPOLOGY_TYPE);
 }
 
-static int skx_iio_set_mapping(struct intel_uncore_type *type)
+static void skx_iio_set_mapping(struct intel_uncore_type *type)
 {
-	return pmu_iio_set_mapping(type, &skx_iio_mapping_group);
+	pmu_iio_set_mapping(type, &skx_iio_mapping_group);
 }
 
 static void skx_iio_cleanup_mapping(struct intel_uncore_type *type)
 {
-	pmu_iio_cleanup_mapping(type, &skx_iio_mapping_group);
+	pmu_cleanup_mapping(type, &skx_iio_mapping_group);
 }
 
 static struct intel_uncore_type skx_uncore_iio = {
@@ -4139,6 +4270,132 @@ static struct intel_uncore_ops skx_upi_uncore_pci_ops = {
 	.read_counter	= snbep_uncore_pci_read_counter,
 };
 
+static umode_t
+skx_upi_mapping_visible(struct kobject *kobj, struct attribute *attr, int die)
+{
+	struct intel_uncore_pmu *pmu = dev_to_uncore_pmu(kobj_to_dev(kobj));
+
+	return pmu->type->topology[die][pmu->pmu_idx].upi->enabled ? attr->mode : 0;
+}
+
+static ssize_t skx_upi_mapping_show(struct device *dev,
+				    struct device_attribute *attr, char *buf)
+{
+	struct intel_uncore_pmu *pmu = dev_to_uncore_pmu(dev);
+	struct dev_ext_attribute *ea = to_dev_ext_attribute(attr);
+	long die = (long)ea->var;
+	struct uncore_upi_topology *upi = pmu->type->topology[die][pmu->pmu_idx].upi;
+
+	return sysfs_emit(buf, "upi_%d,die_%d\n", upi->pmu_idx_to, upi->die_to);
+}
+
+#define SKX_UPI_REG_DID			0x2058
+#define SKX_UPI_REGS_ADDR_DEVICE_LINK0	0x0e
+#define SKX_UPI_REGS_ADDR_FUNCTION	0x00
+
+/*
+ * UPI Link Parameter 0
+ * |  Bit  |  Default  |  Description
+ * | 19:16 |     0h    | base_nodeid - The NodeID of the sending socket.
+ * | 12:8  |    00h    | sending_port - The processor die port number of the sending port.
+ */
+#define SKX_KTILP0_OFFSET	0x94
+
+/*
+ * UPI Pcode Status. This register is used by PCode to store the link training status.
+ * |  Bit  |  Default  |  Description
+ * |   4   |     0h    | ll_status_valid — Bit indicates the valid training status
+ *                       logged from PCode to the BIOS.
+ */
+#define SKX_KTIPCSTS_OFFSET	0x120
+
+static int upi_fill_topology(struct pci_dev *dev, struct intel_uncore_topology *tp,
+			     int pmu_idx)
+{
+	int ret;
+	u32 upi_conf;
+	struct uncore_upi_topology *upi = tp->upi;
+
+	tp->pmu_idx = pmu_idx;
+	ret = pci_read_config_dword(dev, SKX_KTIPCSTS_OFFSET, &upi_conf);
+	if (ret) {
+		ret = pcibios_err_to_errno(ret);
+		goto err;
+	}
+	upi->enabled = (upi_conf >> 4) & 1;
+	if (upi->enabled) {
+		ret = pci_read_config_dword(dev, SKX_KTILP0_OFFSET,
+					    &upi_conf);
+		if (ret) {
+			ret = pcibios_err_to_errno(ret);
+			goto err;
+		}
+		upi->die_to = (upi_conf >> 16) & 0xf;
+		upi->pmu_idx_to = (upi_conf >> 8) & 0x1f;
+	}
+err:
+	return ret;
+}
+
+static int skx_upi_topology_cb(struct intel_uncore_type *type, int segment,
+				int die, u64 cpu_bus_msr)
+{
+	int idx, ret;
+	struct intel_uncore_topology *upi;
+	unsigned int devfn;
+	struct pci_dev *dev = NULL;
+	u8 bus = cpu_bus_msr >> (3 * BUS_NUM_STRIDE);
+
+	for (idx = 0; idx < type->num_boxes; idx++) {
+		upi = &type->topology[die][idx];
+		devfn = PCI_DEVFN(SKX_UPI_REGS_ADDR_DEVICE_LINK0 + idx,
+				  SKX_UPI_REGS_ADDR_FUNCTION);
+		dev = pci_get_domain_bus_and_slot(segment, bus, devfn);
+		if (dev) {
+			ret = upi_fill_topology(dev, upi, idx);
+			if (ret)
+				break;
+		}
+	}
+
+	pci_dev_put(dev);
+	return ret;
+}
+
+static int skx_upi_get_topology(struct intel_uncore_type *type)
+{
+	/* CPX case is not supported */
+	if (boot_cpu_data.x86_stepping == 11)
+		return -EPERM;
+
+	return skx_pmu_get_topology(type, skx_upi_topology_cb);
+}
+
+static struct attribute_group skx_upi_mapping_group = {
+	.is_visible	= skx_upi_mapping_visible,
+};
+
+static const struct attribute_group *skx_upi_attr_update[] = {
+	&skx_upi_mapping_group,
+	NULL
+};
+
+static void
+pmu_upi_set_mapping(struct intel_uncore_type *type, struct attribute_group *ag)
+{
+	pmu_set_mapping(type, ag, skx_upi_mapping_show, UPI_TOPOLOGY_TYPE);
+}
+
+static void skx_upi_set_mapping(struct intel_uncore_type *type)
+{
+	pmu_upi_set_mapping(type, &skx_upi_mapping_group);
+}
+
+static void skx_upi_cleanup_mapping(struct intel_uncore_type *type)
+{
+	pmu_cleanup_mapping(type, &skx_upi_mapping_group);
+}
+
 static struct intel_uncore_type skx_uncore_upi = {
 	.name		= "upi",
 	.num_counters   = 4,
@@ -4151,6 +4408,10 @@ static struct intel_uncore_type skx_uncore_upi = {
 	.box_ctl	= SKX_UPI_PCI_PMON_BOX_CTL,
 	.ops		= &skx_upi_uncore_pci_ops,
 	.format_group	= &skx_upi_uncore_format_group,
+	.attr_update	= skx_upi_attr_update,
+	.get_topology	= skx_upi_get_topology,
+	.set_mapping	= skx_upi_set_mapping,
+	.cleanup_mapping = skx_upi_cleanup_mapping,
 };
 
 static void skx_m2m_uncore_pci_init_box(struct intel_uncore_box *box)
@@ -4461,11 +4722,6 @@ static int sad_cfg_iio_topology(struct intel_uncore_type *type, u8 *sad_pmon_map
 	int die, stack_id, ret = -EPERM;
 	struct pci_dev *dev = NULL;
 
-	type->topology = kcalloc(uncore_max_dies(), sizeof(*type->topology),
-				 GFP_KERNEL);
-	if (!type->topology)
-		return -ENOMEM;
-
 	while ((dev = pci_get_device(PCI_VENDOR_ID_INTEL, SNR_ICX_MESH2IIO_MMAP_DID, dev))) {
 		ret = pci_read_config_dword(dev, SNR_ICX_SAD_CONTROL_CFG, &sad_cfg);
 		if (ret) {
@@ -4483,14 +4739,12 @@ static int sad_cfg_iio_topology(struct intel_uncore_type *type, u8 *sad_pmon_map
 		/* Convert stack id from SAD_CONTROL to PMON notation. */
 		stack_id = sad_pmon_mapping[stack_id];
 
-		((u8 *)&(type->topology[die].configuration))[stack_id] = dev->bus->number;
-		type->topology[die].segment = pci_domain_nr(dev->bus);
+		type->topology[die][stack_id].iio->segment = pci_domain_nr(dev->bus);
+		type->topology[die][stack_id].pmu_idx = stack_id;
+		type->topology[die][stack_id].iio->pci_bus_no = dev->bus->number;
 	}
 
-	if (ret) {
-		kfree(type->topology);
-		type->topology = NULL;
-	}
+	pci_dev_put(dev);
 
 	return ret;
 }
@@ -4519,14 +4773,14 @@ static int snr_iio_get_topology(struct intel_uncore_type *type)
 	return sad_cfg_iio_topology(type, snr_sad_pmon_mapping);
 }
 
-static int snr_iio_set_mapping(struct intel_uncore_type *type)
+static void snr_iio_set_mapping(struct intel_uncore_type *type)
 {
-	return pmu_iio_set_mapping(type, &snr_iio_mapping_group);
+	pmu_iio_set_mapping(type, &snr_iio_mapping_group);
 }
 
 static void snr_iio_cleanup_mapping(struct intel_uncore_type *type)
 {
-	pmu_iio_cleanup_mapping(type, &snr_iio_mapping_group);
+	pmu_cleanup_mapping(type, &snr_iio_mapping_group);
 }
 
 static struct event_constraint snr_uncore_iio_constraints[] = {
@@ -4857,6 +5111,8 @@ static int snr_uncore_mmio_map(struct intel_uncore_box *box,
 
 	addr += box_ctl;
 
+	pci_dev_put(pdev);
+
 	box->io_addr = ioremap(addr, type->mmio_map_size);
 	if (!box->io_addr) {
 		pr_warn("perf uncore: Failed to ioremap for %s.\n", type->name);
@@ -5137,14 +5393,19 @@ static int icx_iio_get_topology(struct intel_uncore_type *type)
 	return sad_cfg_iio_topology(type, icx_sad_pmon_mapping);
 }
 
-static int icx_iio_set_mapping(struct intel_uncore_type *type)
+static void icx_iio_set_mapping(struct intel_uncore_type *type)
 {
-	return pmu_iio_set_mapping(type, &icx_iio_mapping_group);
+	/* Detect ICX-D system. This case is not supported */
+	if (boot_cpu_data.x86_model == INTEL_FAM6_ICELAKE_D) {
+		pmu_clear_mapping_attr(type->attr_update, &icx_iio_mapping_group);
+		return;
+	}
+	pmu_iio_set_mapping(type, &icx_iio_mapping_group);
 }
 
 static void icx_iio_cleanup_mapping(struct intel_uncore_type *type)
 {
-	pmu_iio_cleanup_mapping(type, &icx_iio_mapping_group);
+	pmu_cleanup_mapping(type, &icx_iio_mapping_group);
 }
 
 static struct intel_uncore_type icx_uncore_iio = {
@@ -5337,6 +5598,76 @@ static const struct attribute_group icx_upi_uncore_format_group = {
 	.attrs = icx_upi_uncore_formats_attr,
 };
 
+#define ICX_UPI_REGS_ADDR_DEVICE_LINK0	0x02
+#define ICX_UPI_REGS_ADDR_FUNCTION	0x01
+
+static int discover_upi_topology(struct intel_uncore_type *type, int ubox_did, int dev_link0)
+{
+	struct pci_dev *ubox = NULL;
+	struct pci_dev *dev = NULL;
+	u32 nid, gid;
+	int i, idx, ret = -EPERM;
+	struct intel_uncore_topology *upi;
+	unsigned int devfn;
+
+	/* GIDNIDMAP method supports machines which have less than 8 sockets. */
+	if (uncore_max_dies() > 8)
+		goto err;
+
+	while ((ubox = pci_get_device(PCI_VENDOR_ID_INTEL, ubox_did, ubox))) {
+		ret = upi_nodeid_groupid(ubox, SKX_CPUNODEID, SKX_GIDNIDMAP, &nid, &gid);
+		if (ret) {
+			ret = pcibios_err_to_errno(ret);
+			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;
+				}
+			}
+		}
+	}
+err:
+	pci_dev_put(ubox);
+	pci_dev_put(dev);
+	return ret;
+}
+
+static int icx_upi_get_topology(struct intel_uncore_type *type)
+{
+	return discover_upi_topology(type, ICX_UBOX_DID, ICX_UPI_REGS_ADDR_DEVICE_LINK0);
+}
+
+static struct attribute_group icx_upi_mapping_group = {
+	.is_visible	= skx_upi_mapping_visible,
+};
+
+static const struct attribute_group *icx_upi_attr_update[] = {
+	&icx_upi_mapping_group,
+	NULL
+};
+
+static void icx_upi_set_mapping(struct intel_uncore_type *type)
+{
+	pmu_upi_set_mapping(type, &icx_upi_mapping_group);
+}
+
+static void icx_upi_cleanup_mapping(struct intel_uncore_type *type)
+{
+	pmu_cleanup_mapping(type, &icx_upi_mapping_group);
+}
+
 static struct intel_uncore_type icx_uncore_upi = {
 	.name		= "upi",
 	.num_counters   = 4,
@@ -5349,6 +5680,10 @@ static struct intel_uncore_type icx_uncore_upi = {
 	.box_ctl	= ICX_UPI_PCI_PMON_BOX_CTL,
 	.ops		= &skx_upi_uncore_pci_ops,
 	.format_group	= &icx_upi_uncore_format_group,
+	.attr_update	= icx_upi_attr_update,
+	.get_topology	= icx_upi_get_topology,
+	.set_mapping	= icx_upi_set_mapping,
+	.cleanup_mapping = icx_upi_cleanup_mapping,
 };
 
 static struct event_constraint icx_uncore_m3upi_constraints[] = {
@@ -5780,9 +6115,43 @@ static struct intel_uncore_type spr_uncore_m2m = {
 	.name			= "m2m",
 };
 
+static struct attribute_group spr_upi_mapping_group = {
+	.is_visible	= skx_upi_mapping_visible,
+};
+
+static const struct attribute_group *spr_upi_attr_update[] = {
+	&uncore_alias_group,
+	&spr_upi_mapping_group,
+	NULL
+};
+
+#define SPR_UPI_REGS_ADDR_DEVICE_LINK0	0x01
+
+static void spr_upi_set_mapping(struct intel_uncore_type *type)
+{
+	pmu_upi_set_mapping(type, &spr_upi_mapping_group);
+}
+
+static void spr_upi_cleanup_mapping(struct intel_uncore_type *type)
+{
+	pmu_cleanup_mapping(type, &spr_upi_mapping_group);
+}
+
+static int spr_upi_get_topology(struct intel_uncore_type *type)
+{
+	return discover_upi_topology(type, SPR_UBOX_DID, SPR_UPI_REGS_ADDR_DEVICE_LINK0);
+}
+
 static struct intel_uncore_type spr_uncore_upi = {
-	SPR_UNCORE_PCI_COMMON_FORMAT(),
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.event_mask_ext		= SPR_RAW_EVENT_MASK_EXT,
+	.format_group		= &spr_uncore_raw_format_group,
+	.ops			= &spr_uncore_pci_ops,
 	.name			= "upi",
+	.attr_update		= spr_upi_attr_update,
+	.get_topology		= spr_upi_get_topology,
+	.set_mapping		= spr_upi_set_mapping,
+	.cleanup_mapping	= spr_upi_cleanup_mapping,
 };
 
 static struct intel_uncore_type spr_uncore_m3upi = {
@@ -5986,6 +6355,12 @@ static void uncore_type_customized_copy(struct intel_uncore_type *to_type,
 		to_type->format_group = from_type->format_group;
 	if (from_type->attr_update)
 		to_type->attr_update = from_type->attr_update;
+	if (from_type->set_mapping)
+		to_type->set_mapping = from_type->set_mapping;
+	if (from_type->get_topology)
+		to_type->get_topology = from_type->get_topology;
+	if (from_type->cleanup_mapping)
+		to_type->cleanup_mapping = from_type->cleanup_mapping;
 }
 
 static struct intel_uncore_type **