Merge branch 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull perf updates from Ingo Molnar:
 "Kernel side changes:

   - Improved kbprobes robustness

   - Intel PEBS support for PT hardware tracing

   - Other Intel PT improvements: high order pages memory footprint
     reduction and various related cleanups

   - Misc cleanups

  The perf tooling side has been very busy in this cycle, with over 300
  commits. This is an incomplete high-level summary of the many
  improvements done by over 30 developers:

   - Lots of updates to the following tools:

      'perf c2c'
      'perf config'
      'perf record'
      'perf report'
      'perf script'
      'perf test'
      'perf top'
      'perf trace'

   - Updates to libperf and libtraceevent, and a consolidation of the
     proliferation of x86 instruction decoder libraries.

   - Vendor event updates for Intel and PowerPC CPUs,

   - Updates to hardware tracing tooling for ARM and Intel CPUs,

   - ... and lots of other changes and cleanups - see the shortlog and
     Git log for details"

* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (322 commits)
  kprobes: Prohibit probing on BUG() and WARN() address
  perf/x86: Make more stuff static
  x86, perf: Fix the dependency of the x86 insn decoder selftest
  objtool: Ignore intentional differences for the x86 insn decoder
  objtool: Update sync-check.sh from perf's check-headers.sh
  perf build: Ignore intentional differences for the x86 insn decoder
  perf intel-pt: Use shared x86 insn decoder
  perf intel-pt: Remove inat.c from build dependency list
  perf: Update .gitignore file
  objtool: Move x86 insn decoder to a common location
  perf metricgroup: Support multiple events for metricgroup
  perf metricgroup: Scale the metric result
  perf pmu: Change convert_scale from static to global
  perf symbols: Move mem_info and branch_info out of symbol.h
  perf auxtrace: Uninline functions that touch perf_session
  perf tools: Remove needless evlist.h include directives
  perf tools: Remove needless evlist.h include directives
  perf tools: Remove needless thread_map.h include directives
  perf tools: Remove needless thread.h include directives
  perf tools: Remove needless map.h include directives
  ...

1 files changed, 93 insertions, 0 deletions

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 0463c1151bae..2aad959e6def 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -1887,6 +1887,89 @@ list_del_event(struct perf_event *event, struct perf_event_context *ctx)
 	ctx->generation++;
 }
 
+static int
+perf_aux_output_match(struct perf_event *event, struct perf_event *aux_event)
+{
+	if (!has_aux(aux_event))
+		return 0;
+
+	if (!event->pmu->aux_output_match)
+		return 0;
+
+	return event->pmu->aux_output_match(aux_event);
+}
+
+static void put_event(struct perf_event *event);
+static void event_sched_out(struct perf_event *event,
+			    struct perf_cpu_context *cpuctx,
+			    struct perf_event_context *ctx);
+
+static void perf_put_aux_event(struct perf_event *event)
+{
+	struct perf_event_context *ctx = event->ctx;
+	struct perf_cpu_context *cpuctx = __get_cpu_context(ctx);
+	struct perf_event *iter;
+
+	/*
+	 * If event uses aux_event tear down the link
+	 */
+	if (event->aux_event) {
+		iter = event->aux_event;
+		event->aux_event = NULL;
+		put_event(iter);
+		return;
+	}
+
+	/*
+	 * If the event is an aux_event, tear down all links to
+	 * it from other events.
+	 */
+	for_each_sibling_event(iter, event->group_leader) {
+		if (iter->aux_event != event)
+			continue;
+
+		iter->aux_event = NULL;
+		put_event(event);
+
+		/*
+		 * If it's ACTIVE, schedule it out and put it into ERROR
+		 * state so that we don't try to schedule it again. Note
+		 * that perf_event_enable() will clear the ERROR status.
+		 */
+		event_sched_out(iter, cpuctx, ctx);
+		perf_event_set_state(event, PERF_EVENT_STATE_ERROR);
+	}
+}
+
+static int perf_get_aux_event(struct perf_event *event,
+			      struct perf_event *group_leader)
+{
+	/*
+	 * Our group leader must be an aux event if we want to be
+	 * an aux_output. This way, the aux event will precede its
+	 * aux_output events in the group, and therefore will always
+	 * schedule first.
+	 */
+	if (!group_leader)
+		return 0;
+
+	if (!perf_aux_output_match(event, group_leader))
+		return 0;
+
+	if (!atomic_long_inc_not_zero(&group_leader->refcount))
+		return 0;
+
+	/*
+	 * Link aux_outputs to their aux event; this is undone in
+	 * perf_group_detach() by perf_put_aux_event(). When the
+	 * group in torn down, the aux_output events loose their
+	 * link to the aux_event and can't schedule any more.
+	 */
+	event->aux_event = group_leader;
+
+	return 1;
+}
+
 static void perf_group_detach(struct perf_event *event)
 {
 	struct perf_event *sibling, *tmp;
@@ -1902,6 +1985,8 @@ static void perf_group_detach(struct perf_event *event)
 
 	event->attach_state &= ~PERF_ATTACH_GROUP;
 
+	perf_put_aux_event(event);
+
 	/*
 	 * If this is a sibling, remove it from its group.
 	 */
@@ -10426,6 +10511,12 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu,
 		goto err_ns;
 	}
 
+	if (event->attr.aux_output &&
+	    !(pmu->capabilities & PERF_PMU_CAP_AUX_OUTPUT)) {
+		err = -EOPNOTSUPP;
+		goto err_pmu;
+	}
+
 	err = exclusive_event_init(event);
 	if (err)
 		goto err_pmu;
@@ -11082,6 +11173,8 @@ SYSCALL_DEFINE5(perf_event_open,
 		}
 	}
 
+	if (event->attr.aux_output && !perf_get_aux_event(event, group_leader))
+		goto err_locked;
 
 	/*
 	 * Must be under the same ctx::mutex as perf_install_in_context(),