From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752010Ab3LXIWl (ORCPT ); Tue, 24 Dec 2013 03:22:41 -0500 Received: from LGEMRELSE6Q.lge.com ([156.147.1.121]:47871 "EHLO LGEMRELSE6Q.lge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751706Ab3LXIWe (ORCPT ); Tue, 24 Dec 2013 03:22:34 -0500 X-AuditID: 9c930179-b7c89ae000006438-f1-52b944470183 From: Namhyung Kim To: Arnaldo Carvalho de Melo Cc: Peter Zijlstra , Paul Mackerras , Ingo Molnar , Namhyung Kim , LKML , Arun Sharma , Frederic Weisbecker , Jiri Olsa , Rodrigo Campos Subject: [PATCH 07/21] perf report: Cache cumulative callchains Date: Tue, 24 Dec 2013 17:22:13 +0900 Message-Id: <1387873347-28838-8-git-send-email-namhyung@kernel.org> X-Mailer: git-send-email 1.7.11.7 In-Reply-To: <1387873347-28838-1-git-send-email-namhyung@kernel.org> References: <1387873347-28838-1-git-send-email-namhyung@kernel.org> X-Brightmail-Tracker: AAAAAA== Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Namhyung Kim It is possble that a callchain has cycles or recursive calls. In that case it'll end up having entries more than 100% overhead in the output. In order to prevent such entries, cache each callchain node and skip if same entry already cumulated. Cc: Arun Sharma Cc: Frederic Weisbecker Signed-off-by: Namhyung Kim --- tools/perf/builtin-report.c | 48 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 48 insertions(+) diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 17c41c686042..1314841ffad1 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -368,8 +368,27 @@ iter_prepare_cumulative_entry(struct add_entry_iter *iter, struct addr_location *al __maybe_unused, struct perf_sample *sample) { + struct callchain_cursor_node *node; + struct hist_entry **he_cache; + callchain_cursor_commit(&callchain_cursor); + /* + * This is for detecting cycles or recursions so that they're + * cumulated only one time to prevent entries more than 100% + * overhead. + */ + he_cache = malloc(sizeof(*he_cache) * (PERF_MAX_STACK_DEPTH + 1)); + if (he_cache == NULL) + return -ENOMEM; + + iter->priv = he_cache; + iter->curr = 0; + + node = callchain_cursor_current(&callchain_cursor); + if (node == NULL) + return 0; + iter->evsel = evsel; iter->sample = sample; iter->machine = machine; @@ -382,6 +401,7 @@ iter_add_single_cumulative_entry(struct add_entry_iter *iter, { struct perf_evsel *evsel = iter->evsel; struct perf_sample *sample = iter->sample; + struct hist_entry **he_cache = iter->priv; struct hist_entry *he; he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL, @@ -390,6 +410,8 @@ iter_add_single_cumulative_entry(struct add_entry_iter *iter, if (he == NULL) return -ENOMEM; + he_cache[iter->curr++] = he; + /* * This is for putting parents upward during output resort iff * only a child gets sampled. See hist_entry__sort_on_period(). @@ -455,7 +477,29 @@ iter_add_next_cumulative_entry(struct add_entry_iter *iter, { struct perf_evsel *evsel = iter->evsel; struct perf_sample *sample = iter->sample; + struct hist_entry **he_cache = iter->priv; struct hist_entry *he; + struct hist_entry he_tmp = { + .cpu = al->cpu, + .thread = al->thread, + .comm = thread__comm(al->thread), + .ip = al->addr, + .ms = { + .map = al->map, + .sym = al->sym, + }, + .parent = iter->parent, + }; + int i; + + /* + * Check if there's duplicate entries in the callchain. + * It's possible that it has cycles or recursive calls. + */ + for (i = 0; i < iter->curr; i++) { + if (hist_entry__cmp(he_cache[i], &he_tmp) == 0) + return 0; + } he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL, sample->period, sample->weight, @@ -463,6 +507,8 @@ iter_add_next_cumulative_entry(struct add_entry_iter *iter, if (he == NULL) return -ENOMEM; + he_cache[iter->curr++] = he; + return hist_entry__inc_addr_samples(he, evsel->idx, al->addr); } @@ -476,6 +522,8 @@ iter_finish_cumulative_entry(struct add_entry_iter *iter, evsel->hists.stats.total_period += sample->period; hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE); + free(iter->priv); + iter->priv = NULL; return 0; } -- 1.7.11.7