perf: Factor out auxiliary events notification

Add perf_event_aux() function to send out all types of
auxiliary events - mmap, task, comm events. For each type
there's match and output functions defined and used as
callbacks during perf_event_aux processing.

This way we can centralize the pmu/context iterating and
event matching logic. Also since lot of the code was
duplicated, this patch reduces the .text size about 2kB
on my setup:

  snipped output from 'objdump -x kernel/events/core.o'

  before:
  Idx Name          Size
    0 .text         0000d313

  after:
  Idx Name          Size
    0 .text         0000cad3

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: Borislav Petkov <bp@alien8.de>
Link: http://lkml.kernel.org/r/1367857638-27631-3-git-send-email-jolsa@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>

1 files changed, 89 insertions, 153 deletions

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 38b68a05c3c6..9dc297faf7c0 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -4394,6 +4394,64 @@ perf_event_read_event(struct perf_event *event,
         perf_output_end(&handle);
 }
 
+typedef int  (perf_event_aux_match_cb)(struct perf_event *event, void *data);
+typedef void (perf_event_aux_output_cb)(struct perf_event *event, void *data);
+
+static void
+perf_event_aux_ctx(struct perf_event_context *ctx,
+                   perf_event_aux_match_cb match,
+                   perf_event_aux_output_cb output,
+                   void *data)
+{
+        struct perf_event *event;
+
+        list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
+                if (event->state < PERF_EVENT_STATE_INACTIVE)
+                        continue;
+                if (!event_filter_match(event))
+                        continue;
+                if (match(event, data))
+                        output(event, data);
+        }
+}
+
+static void
+perf_event_aux(perf_event_aux_match_cb match,
+               perf_event_aux_output_cb output,
+               void *data,
+               struct perf_event_context *task_ctx)
+{
+        struct perf_cpu_context *cpuctx;
+        struct perf_event_context *ctx;
+        struct pmu *pmu;
+        int ctxn;
+
+        rcu_read_lock();
+        list_for_each_entry_rcu(pmu, &pmus, entry) {
+                cpuctx = get_cpu_ptr(pmu->pmu_cpu_context);
+                if (cpuctx->unique_pmu != pmu)
+                        goto next;
+                perf_event_aux_ctx(&cpuctx->ctx, match, output, data);
+                if (task_ctx)
+                        goto next;
+                ctxn = pmu->task_ctx_nr;
+                if (ctxn < 0)
+                        goto next;
+                ctx = rcu_dereference(current->perf_event_ctxp[ctxn]);
+                if (ctx)
+                        perf_event_aux_ctx(ctx, match, output, data);
+next:
+                put_cpu_ptr(pmu->pmu_cpu_context);
+        }
+
+        if (task_ctx) {
+                preempt_disable();
+                perf_event_aux_ctx(task_ctx, match, output, data);
+                preempt_enable();
+        }
+        rcu_read_unlock();
+}
+
 /*
  * task tracking -- fork/exit
  *
@@ -4416,8 +4474,9 @@ struct perf_task_event {
 };
 
 static void perf_event_task_output(struct perf_event *event,
-                                     struct perf_task_event *task_event)
+                                   void *data)
 {
+        struct perf_task_event *task_event = data;
         struct perf_output_handle handle;
         struct perf_sample_data sample;
         struct task_struct *task = task_event->task;
@@ -4445,64 +4504,11 @@ out:
         task_event->event_id.header.size = size;
 }
 
-static int perf_event_task_match(struct perf_event *event)
-{
-        if (event->state < PERF_EVENT_STATE_INACTIVE)
-                return 0;
-
-        if (!event_filter_match(event))
-                return 0;
-
-        if (event->attr.comm || event->attr.mmap ||
-            event->attr.mmap_data || event->attr.task)
-                return 1;
-
-        return 0;
-}
-
-static void perf_event_task_ctx(struct perf_event_context *ctx,
-                                  struct perf_task_event *task_event)
-{
-        struct perf_event *event;
-
-        list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
-                if (perf_event_task_match(event))
-                        perf_event_task_output(event, task_event);
-        }
-}
-
-static void perf_event_task_event(struct perf_task_event *task_event)
+static int perf_event_task_match(struct perf_event *event,
+                                 void *data __maybe_unused)
 {
-        struct perf_cpu_context *cpuctx;
-        struct perf_event_context *ctx, *task_ctx = task_event->task_ctx;
-        struct pmu *pmu;
-        int ctxn;
-
-        rcu_read_lock();
-        list_for_each_entry_rcu(pmu, &pmus, entry) {
-                cpuctx = get_cpu_ptr(pmu->pmu_cpu_context);
-                if (cpuctx->unique_pmu != pmu)
-                        goto next;
-                perf_event_task_ctx(&cpuctx->ctx, task_event);
-
-                if (task_ctx)
-                        goto next;
-                ctxn = pmu->task_ctx_nr;
-                if (ctxn < 0)
-                        goto next;
-                ctx = rcu_dereference(current->perf_event_ctxp[ctxn]);
-                if (ctx)
-                        perf_event_task_ctx(ctx, task_event);
-next:
-                put_cpu_ptr(pmu->pmu_cpu_context);
-        }
-        if (task_ctx) {
-                preempt_disable();
-                perf_event_task_ctx(task_ctx, task_event);
-                preempt_enable();
-        }
-
-        rcu_read_unlock();
+        return event->attr.comm || event->attr.mmap ||
+               event->attr.mmap_data || event->attr.task;
 }
 
 static void perf_event_task(struct task_struct *task,
@@ -4533,7 +4539,10 @@ static void perf_event_task(struct task_struct *task,
                 },
         };
 
-        perf_event_task_event(&task_event);
+        perf_event_aux(perf_event_task_match,
+                       perf_event_task_output,
+                       &task_event,
+                       task_ctx);
 }
 
 void perf_event_fork(struct task_struct *task)
@@ -4559,8 +4568,9 @@ struct perf_comm_event {
 };
 
 static void perf_event_comm_output(struct perf_event *event,
-                                     struct perf_comm_event *comm_event)
+                                   void *data)
 {
+        struct perf_comm_event *comm_event = data;
         struct perf_output_handle handle;
         struct perf_sample_data sample;
         int size = comm_event->event_id.header.size;
@@ -4587,39 +4597,16 @@ out:
         comm_event->event_id.header.size = size;
 }
 
-static int perf_event_comm_match(struct perf_event *event)
-{
-        if (event->state < PERF_EVENT_STATE_INACTIVE)
-                return 0;
-
-        if (!event_filter_match(event))
-                return 0;
-
-        if (event->attr.comm)
-                return 1;
-
-        return 0;
-}
-
-static void perf_event_comm_ctx(struct perf_event_context *ctx,
-                                  struct perf_comm_event *comm_event)
+static int perf_event_comm_match(struct perf_event *event,
+                                 void *data __maybe_unused)
 {
-        struct perf_event *event;
-
-        list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
-                if (perf_event_comm_match(event))
-                        perf_event_comm_output(event, comm_event);
-        }
+        return event->attr.comm;
 }
 
 static void perf_event_comm_event(struct perf_comm_event *comm_event)
 {
-        struct perf_cpu_context *cpuctx;
-        struct perf_event_context *ctx;
         char comm[TASK_COMM_LEN];
         unsigned int size;
-        struct pmu *pmu;
-        int ctxn;
 
         memset(comm, 0, sizeof(comm));
         strlcpy(comm, comm_event->task->comm, sizeof(comm));
@@ -4629,24 +4616,11 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event)
         comm_event->comm_size = size;
 
         comm_event->event_id.header.size = sizeof(comm_event->event_id) + size;
-        rcu_read_lock();
-        list_for_each_entry_rcu(pmu, &pmus, entry) {
-                cpuctx = get_cpu_ptr(pmu->pmu_cpu_context);
-                if (cpuctx->unique_pmu != pmu)
-                        goto next;
-                perf_event_comm_ctx(&cpuctx->ctx, comm_event);
 
-                ctxn = pmu->task_ctx_nr;
-                if (ctxn < 0)
-                        goto next;
-
-                ctx = rcu_dereference(current->perf_event_ctxp[ctxn]);
-                if (ctx)
-                        perf_event_comm_ctx(ctx, comm_event);
-next:
-                put_cpu_ptr(pmu->pmu_cpu_context);
-        }
-        rcu_read_unlock();
+        perf_event_aux(perf_event_comm_match,
+                       perf_event_comm_output,
+                       comm_event,
+                       NULL);
 }
 
 void perf_event_comm(struct task_struct *task)
@@ -4708,8 +4682,9 @@ struct perf_mmap_event {
 };
 
 static void perf_event_mmap_output(struct perf_event *event,
-                                     struct perf_mmap_event *mmap_event)
+                                   void *data)
 {
+        struct perf_mmap_event *mmap_event = data;
         struct perf_output_handle handle;
         struct perf_sample_data sample;
         int size = mmap_event->event_id.header.size;
@@ -4736,46 +4711,24 @@ out:
 }
 
 static int perf_event_mmap_match(struct perf_event *event,
-                                   struct perf_mmap_event *mmap_event,
-                                   int executable)
+                                 void *data)
 {
-        if (event->state < PERF_EVENT_STATE_INACTIVE)
-                return 0;
-
-        if (!event_filter_match(event))
-                return 0;
-
-        if ((!executable && event->attr.mmap_data) ||
-            (executable && event->attr.mmap))
-                return 1;
-
-        return 0;
-}
-
-static void perf_event_mmap_ctx(struct perf_event_context *ctx,
-                                  struct perf_mmap_event *mmap_event,
-                                  int executable)
-{
-        struct perf_event *event;
+        struct perf_mmap_event *mmap_event = data;
+        struct vm_area_struct *vma = mmap_event->vma;
+        int executable = vma->vm_flags & VM_EXEC;
 
-        list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
-                if (perf_event_mmap_match(event, mmap_event, executable))
-                        perf_event_mmap_output(event, mmap_event);
-        }
+        return (!executable && event->attr.mmap_data) ||
+               (executable && event->attr.mmap);
 }
 
 static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
 {
-        struct perf_cpu_context *cpuctx;
-        struct perf_event_context *ctx;
         struct vm_area_struct *vma = mmap_event->vma;
         struct file *file = vma->vm_file;
         unsigned int size;
         char tmp[16];
         char *buf = NULL;
         const char *name;
-        struct pmu *pmu;
-        int ctxn;
 
         memset(tmp, 0, sizeof(tmp));
 
@@ -4831,27 +4784,10 @@ got_name:
 
         mmap_event->event_id.header.size = sizeof(mmap_event->event_id) + size;
 
-        rcu_read_lock();
-        list_for_each_entry_rcu(pmu, &pmus, entry) {
-                cpuctx = get_cpu_ptr(pmu->pmu_cpu_context);
-                if (cpuctx->unique_pmu != pmu)
-                        goto next;
-                perf_event_mmap_ctx(&cpuctx->ctx, mmap_event,
-                                        vma->vm_flags & VM_EXEC);
-
-                ctxn = pmu->task_ctx_nr;
-                if (ctxn < 0)
-                        goto next;
-
-                ctx = rcu_dereference(current->perf_event_ctxp[ctxn]);
-                if (ctx) {
-                        perf_event_mmap_ctx(ctx, mmap_event,
-                                        vma->vm_flags & VM_EXEC);
-                }
-next:
-                put_cpu_ptr(pmu->pmu_cpu_context);
-        }
-        rcu_read_unlock();
+        perf_event_aux(perf_event_mmap_match,
+                       perf_event_mmap_output,
+                       mmap_event,
+                       NULL);
 
         kfree(buf);
 }