perf/x86: Add new event for AUX output counter index

PEBS-via-PT records contain a mask of applicable counters. To identify
which event belongs to which counter, a side-band event is needed. Until
now, there has been no side-band event, and consequently users were limited
to using a single event.

Add such a side-band event. Note the event is optimised to output only
when the counter index changes for an event. That works only so long as
all PEBS-via-PT events are scheduled together, which they are for a
recording session because they are in a single group.

Also no attribute bit is used to select the new event, so a new
kernel is not compatible with older perf tools.  The assumption
being that PEBS-via-PT is sufficiently esoteric that users will not
be troubled by this.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210907163903.11820-2-adrian.hunter@intel.com

3 files changed, 23 insertions, 0 deletions

diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c
index 2a57dbed4894..be33423e9762 100644
--- a/arch/x86/events/core.c
+++ b/arch/x86/events/core.c
@@ -66,6 +66,8 @@ DEFINE_STATIC_CALL_NULL(x86_pmu_enable_all,  *x86_pmu.enable_all);
 DEFINE_STATIC_CALL_NULL(x86_pmu_enable,	     *x86_pmu.enable);
 DEFINE_STATIC_CALL_NULL(x86_pmu_disable,     *x86_pmu.disable);
 
+DEFINE_STATIC_CALL_NULL(x86_pmu_assign, *x86_pmu.assign);
+
 DEFINE_STATIC_CALL_NULL(x86_pmu_add,  *x86_pmu.add);
 DEFINE_STATIC_CALL_NULL(x86_pmu_del,  *x86_pmu.del);
 DEFINE_STATIC_CALL_NULL(x86_pmu_read, *x86_pmu.read);
@@ -1215,6 +1217,8 @@ static inline void x86_assign_hw_event(struct perf_event *event,
 	hwc->last_cpu = smp_processor_id();
 	hwc->last_tag = ++cpuc->tags[i];
 
+	static_call_cond(x86_pmu_assign)(event, idx);
+
 	switch (hwc->idx) {
 	case INTEL_PMC_IDX_FIXED_BTS:
 	case INTEL_PMC_IDX_FIXED_VLBR:
@@ -2005,6 +2009,8 @@ static void x86_pmu_static_call_update(void)
 	static_call_update(x86_pmu_enable, x86_pmu.enable);
 	static_call_update(x86_pmu_disable, x86_pmu.disable);
 
+	static_call_update(x86_pmu_assign, x86_pmu.assign);
+
 	static_call_update(x86_pmu_add, x86_pmu.add);
 	static_call_update(x86_pmu_del, x86_pmu.del);
 	static_call_update(x86_pmu_read, x86_pmu.read);
diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c
index 7011e87be6d0..a555e7c2dce9 100644
--- a/arch/x86/events/intel/core.c
+++ b/arch/x86/events/intel/core.c
@@ -2402,6 +2402,12 @@ static void intel_pmu_disable_event(struct perf_event *event)
 		intel_pmu_pebs_disable(event);
 }
 
+static void intel_pmu_assign_event(struct perf_event *event, int idx)
+{
+	if (is_pebs_pt(event))
+		perf_report_aux_output_id(event, idx);
+}
+
 static void intel_pmu_del_event(struct perf_event *event)
 {
 	if (needs_branch_stack(event))
@@ -4494,8 +4500,16 @@ static int intel_pmu_check_period(struct perf_event *event, u64 value)
 	return intel_pmu_has_bts_period(event, value) ? -EINVAL : 0;
 }
 
+static void intel_aux_output_init(void)
+{
+	/* Refer also intel_pmu_aux_output_match() */
+	if (x86_pmu.intel_cap.pebs_output_pt_available)
+		x86_pmu.assign = intel_pmu_assign_event;
+}
+
 static int intel_pmu_aux_output_match(struct perf_event *event)
 {
+	/* intel_pmu_assign_event() is needed, refer intel_aux_output_init() */
 	if (!x86_pmu.intel_cap.pebs_output_pt_available)
 		return 0;
 
@@ -6301,6 +6315,8 @@ __init int intel_pmu_init(void)
 	if (is_hybrid())
 		intel_pmu_check_hybrid_pmus((u64)fixed_mask);
 
+	intel_aux_output_init();
+
 	return 0;
 }
 
diff --git a/arch/x86/events/perf_event.h b/arch/x86/events/perf_event.h
index e3ac05c97b5e..76436a55d9ba 100644
--- a/arch/x86/events/perf_event.h
+++ b/arch/x86/events/perf_event.h
@@ -726,6 +726,7 @@ struct x86_pmu {
 	void		(*enable_all)(int added);
 	void		(*enable)(struct perf_event *);
 	void		(*disable)(struct perf_event *);
+	void		(*assign)(struct perf_event *event, int idx);
 	void		(*add)(struct perf_event *);
 	void		(*del)(struct perf_event *);
 	void		(*read)(struct perf_event *event);