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()
  ...

3 files changed, 96 insertions, 33 deletions

diff --git a/include/linux/perf/arm_pmu.h b/include/linux/perf/arm_pmu.h
index 0c15c5b7f801..ef914a600087 100644
--- a/include/linux/perf/arm_pmu.h
+++ b/include/linux/perf/arm_pmu.h
@@ -100,7 +100,7 @@ struct arm_pmu {
 	void		(*stop)(struct arm_pmu *);
 	void		(*reset)(void *);
 	int		(*map_event)(struct perf_event *event);
-	int		(*filter_match)(struct perf_event *event);
+	bool		(*filter)(struct pmu *pmu, int cpu);
 	int		num_events;
 	bool		secure_access; /* 32-bit ARM only */
 #define ARMV8_PMUV3_MAX_COMMON_EVENTS		0x40
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 0031f7b4d9ab..c6a3bac76966 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -266,6 +266,7 @@ struct hw_perf_event {
 };
 
 struct perf_event;
+struct perf_event_pmu_context;
 
 /*
  * Common implementation detail of pmu::{start,commit,cancel}_txn
@@ -308,7 +309,7 @@ struct pmu {
 	int				capabilities;
 
 	int __percpu			*pmu_disable_count;
-	struct perf_cpu_context __percpu *pmu_cpu_context;
+	struct perf_cpu_pmu_context __percpu *cpu_pmu_context;
 	atomic_t			exclusive_cnt; /* < 0: cpu; > 0: tsk */
 	int				task_ctx_nr;
 	int				hrtimer_interval_ms;
@@ -443,7 +444,7 @@ struct pmu {
 	/*
 	 * context-switches callback
 	 */
-	void (*sched_task)		(struct perf_event_context *ctx,
+	void (*sched_task)		(struct perf_event_pmu_context *pmu_ctx,
 					bool sched_in);
 
 	/*
@@ -457,8 +458,8 @@ struct pmu {
 	 * implementation and Perf core context switch handling callbacks for usage
 	 * examples.
 	 */
-	void (*swap_task_ctx)		(struct perf_event_context *prev,
-					 struct perf_event_context *next);
+	void (*swap_task_ctx)		(struct perf_event_pmu_context *prev_epc,
+					 struct perf_event_pmu_context *next_epc);
 					/* optional */
 
 	/*
@@ -522,9 +523,10 @@ struct pmu {
 					/* optional */
 
 	/*
-	 * Filter events for PMU-specific reasons.
+	 * Skip programming this PMU on the given CPU. Typically needed for
+	 * big.LITTLE things.
 	 */
-	int (*filter_match)		(struct perf_event *event); /* optional */
+	bool (*filter)			(struct pmu *pmu, int cpu); /* optional */
 
 	/*
 	 * Check period value for PERF_EVENT_IOC_PERIOD ioctl.
@@ -695,6 +697,11 @@ struct perf_event {
 	int				group_caps;
 
 	struct perf_event		*group_leader;
+	/*
+	 * event->pmu will always point to pmu in which this event belongs.
+	 * Whereas event->pmu_ctx->pmu may point to other pmu when group of
+	 * different pmu events is created.
+	 */
 	struct pmu			*pmu;
 	void				*pmu_private;
 
@@ -720,6 +727,12 @@ struct perf_event {
 	struct hw_perf_event		hw;
 
 	struct perf_event_context	*ctx;
+	/*
+	 * event->pmu_ctx points to perf_event_pmu_context in which the event
+	 * is added. This pmu_ctx can be of other pmu for sw event when that
+	 * sw event is part of a group which also contains non-sw events.
+	 */
+	struct perf_event_pmu_context	*pmu_ctx;
 	atomic_long_t			refcount;
 
 	/*
@@ -812,19 +825,69 @@ struct perf_event {
 #endif /* CONFIG_PERF_EVENTS */
 };
 
+/*
+ *           ,-----------------------[1:n]----------------------.
+ *           V                                                  V
+ * perf_event_context <-[1:n]-> perf_event_pmu_context <--- perf_event
+ *           ^                      ^     |                     |
+ *           `--------[1:n]---------'     `-[n:1]-> pmu <-[1:n]-'
+ *
+ *
+ * struct perf_event_pmu_context  lifetime is refcount based and RCU freed
+ * (similar to perf_event_context). Locking is as if it were a member of
+ * perf_event_context; specifically:
+ *
+ *   modification, both: ctx->mutex && ctx->lock
+ *   reading, either:    ctx->mutex || ctx->lock
+ *
+ * There is one exception to this; namely put_pmu_ctx() isn't always called
+ * with ctx->mutex held; this means that as long as we can guarantee the epc
+ * has events the above rules hold.
+ *
+ * Specificially, sys_perf_event_open()'s group_leader case depends on
+ * ctx->mutex pinning the configuration. Since we hold a reference on
+ * group_leader (through the filedesc) it can't go away, therefore it's
+ * associated pmu_ctx must exist and cannot change due to ctx->mutex.
+ */
+struct perf_event_pmu_context {
+	struct pmu			*pmu;
+	struct perf_event_context       *ctx;
+
+	struct list_head		pmu_ctx_entry;
+
+	struct list_head		pinned_active;
+	struct list_head		flexible_active;
+
+	/* Used to avoid freeing per-cpu perf_event_pmu_context */
+	unsigned int			embedded : 1;
+
+	unsigned int			nr_events;
+
+	atomic_t			refcount; /* event <-> epc */
+	struct rcu_head			rcu_head;
+
+	void				*task_ctx_data; /* pmu specific data */
+	/*
+	 * Set when one or more (plausibly active) event can't be scheduled
+	 * due to pmu overcommit or pmu constraints, except tolerant to
+	 * events not necessary to be active due to scheduling constraints,
+	 * such as cgroups.
+	 */
+	int				rotate_necessary;
+};
 
 struct perf_event_groups {
 	struct rb_root	tree;
 	u64		index;
 };
 
+
 /**
  * struct perf_event_context - event context structure
  *
  * Used as a container for task events and CPU events as well:
  */
 struct perf_event_context {
-	struct pmu			*pmu;
 	/*
 	 * Protect the states of the events in the list,
 	 * nr_active, and the list:
@@ -837,27 +900,21 @@ struct perf_event_context {
 	 */
 	struct mutex			mutex;
 
-	struct list_head		active_ctx_list;
+	struct list_head		pmu_ctx_list;
 	struct perf_event_groups	pinned_groups;
 	struct perf_event_groups	flexible_groups;
 	struct list_head		event_list;
 
-	struct list_head		pinned_active;
-	struct list_head		flexible_active;
-
 	int				nr_events;
-	int				nr_active;
 	int				nr_user;
 	int				is_active;
+
+	int				nr_task_data;
 	int				nr_stat;
 	int				nr_freq;
 	int				rotate_disable;
-	/*
-	 * Set when nr_events != nr_active, except tolerant to events not
-	 * necessary to be active due to scheduling constraints, such as cgroups.
-	 */
-	int				rotate_necessary;
-	refcount_t			refcount;
+
+	refcount_t			refcount; /* event <-> ctx */
 	struct task_struct		*task;
 
 	/*
@@ -878,7 +935,6 @@ struct perf_event_context {
 #ifdef CONFIG_CGROUP_PERF
 	int				nr_cgroups;	 /* cgroup evts */
 #endif
-	void				*task_ctx_data; /* pmu specific data */
 	struct rcu_head			rcu_head;
 
 	/*
@@ -896,12 +952,13 @@ struct perf_event_context {
  */
 #define PERF_NR_CONTEXTS	4
 
-/**
- * struct perf_cpu_context - per cpu event context structure
- */
-struct perf_cpu_context {
-	struct perf_event_context	ctx;
-	struct perf_event_context	*task_ctx;
+struct perf_cpu_pmu_context {
+	struct perf_event_pmu_context	epc;
+	struct perf_event_pmu_context	*task_epc;
+
+	struct list_head		sched_cb_entry;
+	int				sched_cb_usage;
+
 	int				active_oncpu;
 	int				exclusive;
 
@@ -909,16 +966,20 @@ struct perf_cpu_context {
 	struct hrtimer			hrtimer;
 	ktime_t				hrtimer_interval;
 	unsigned int			hrtimer_active;
+};
+
+/**
+ * struct perf_event_cpu_context - per cpu event context structure
+ */
+struct perf_cpu_context {
+	struct perf_event_context	ctx;
+	struct perf_event_context	*task_ctx;
+	int				online;
 
 #ifdef CONFIG_CGROUP_PERF
 	struct perf_cgroup		*cgrp;
-	struct list_head		cgrp_cpuctx_entry;
 #endif
 
-	struct list_head		sched_cb_entry;
-	int				sched_cb_usage;
-
-	int				online;
 	/*
 	 * Per-CPU storage for iterators used in visit_groups_merge. The default
 	 * storage is of size 2 to hold the CPU and any CPU event iterators.
@@ -982,6 +1043,8 @@ perf_cgroup_from_task(struct task_struct *task, struct perf_event_context *ctx)
 
 #ifdef CONFIG_PERF_EVENTS
 
+extern struct perf_event_context *perf_cpu_task_ctx(void);
+
 extern void *perf_aux_output_begin(struct perf_output_handle *handle,
 				   struct perf_event *event);
 extern void perf_aux_output_end(struct perf_output_handle *handle,
@@ -1187,7 +1250,7 @@ static inline int is_software_event(struct perf_event *event)
  */
 static inline int in_software_context(struct perf_event *event)
 {
-	return event->ctx->pmu->task_ctx_nr == perf_sw_context;
+	return event->pmu_ctx->pmu->task_ctx_nr == perf_sw_context;
 }
 
 static inline int is_exclusive_pmu(struct pmu *pmu)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index ffb6eb55cd13..4e03f1dcbe52 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1243,7 +1243,7 @@ struct task_struct {
 	unsigned int			futex_state;
 #endif
 #ifdef CONFIG_PERF_EVENTS
-	struct perf_event_context	*perf_event_ctxp[perf_nr_task_contexts];
+	struct perf_event_context	*perf_event_ctxp;
 	struct mutex			perf_event_mutex;
 	struct list_head		perf_event_list;
 #endif