From patchwork Wed Feb 28 16:12:22 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Wang, Weilin" X-Patchwork-Id: 207914 Return-Path: Delivered-To: ouuuleilei@gmail.com Received: by 2002:a05:7300:a81b:b0:108:e6aa:91d0 with SMTP id bq27csp3448734dyb; Wed, 28 Feb 2024 08:14:48 -0800 (PST) X-Forwarded-Encrypted: i=3; AJvYcCW+IL3+ei414l7pLrxhxNN4qO6ndwbW6fy6U5jSatbEybJbTGUio8ueKF9MViOYmz9Krl+ErtdK35F82SNorYMrvAk8GA== X-Google-Smtp-Source: AGHT+IFdPmPHOB3Mho24GFJ1mq9t0m7yl3Th2UJ12lFZwY0+uAdVC94yaESR9LSZJi32698i1s4q X-Received: by 2002:a05:6402:1b07:b0:565:7b61:4c86 with SMTP id by7-20020a0564021b0700b005657b614c86mr9380977edb.6.1709136888046; Wed, 28 Feb 2024 08:14:48 -0800 (PST) ARC-Seal: i=2; a=rsa-sha256; t=1709136888; cv=pass; d=google.com; s=arc-20160816; b=g1QsWdbW9bX0CZJ5cAgF+svXi4DSexVMiHztA81Eev+y0TdKwYnfKOpQLLn/XTCwkn 0HR6cl51WoQFe8lmQbzUtZHAOAhO7Y73iaxhKFEZHzVeQlYChyu9WH6q2t+u0haZMgEz AK1eXPI5RSSd6y2hw6v8f8NoWkJqZzP0dahMqPTUYmdsR5pr1FeLrltAcjZuA3kiKIJ+ 9UkvVPCZ585tWY1fiP6yU+8CI4SCcd+94Dq6CZWwvEugJjIjXau5ExsZnIRSmagbSIbP 3QE3/PZBBKO/GDiK+vlSMVLRqGF67Aje7Jjev8cBYRgWY2tECBMrjslQq+UQuunhhNMO skqg== ARC-Message-Signature: i=2; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=content-transfer-encoding:mime-version:list-unsubscribe :list-subscribe:list-id:precedence:references:in-reply-to:message-id :date:subject:cc:to:from:dkim-signature; bh=iOabe9kxpjO3C2TGdF2Kdkkvv0RDYklZj1OdBjaDtDo=; fh=D8N1uhCLBRzbbRgEj9aMJ2/kVuNqnMf8W9VxSJOfqPA=; b=BlAiPMtQ7wNRHXWyzs8eEByns04VQNUuxhQzGifaRuSDHunqPOsRZXk73LiGqC8WBk kY6N1vXbqG4I/isR8u3FH0jZVcEIAYuC26SPnH7VeiDUYVxgSYBq+XA6bcxqm7U4myPq giL7a9l7+RWTMry2ICH+qdOs452KkOWmHdL7axuPsU32Hhu1z9XNWsedrG/0OtksRkzT 2tU/q/6mnObEmtxQ9EawTNZn3CKVYpxveW0G3z0rSc3BzDIj1B9KXCAZmop0txJFkEH2 m/MA4x+C8D1lmFr3XvyXusBkBCB/80WE+Ncsh9Oxtd40kBLTazMti1BJIJ7eXCQp7U/E ZK6A==; dara=google.com ARC-Authentication-Results: i=2; mx.google.com; dkim=pass header.i=@intel.com header.s=Intel header.b=HqZ3ndw+; arc=pass (i=1 spf=pass spfdomain=intel.com dkim=pass dkdomain=intel.com dmarc=pass fromdomain=intel.com); spf=pass (google.com: domain of linux-kernel+bounces-85334-ouuuleilei=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) smtp.mailfrom="linux-kernel+bounces-85334-ouuuleilei=gmail.com@vger.kernel.org"; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=intel.com Received: from am.mirrors.kernel.org (am.mirrors.kernel.org. [147.75.80.249]) by mx.google.com with ESMTPS id d10-20020a056402400a00b005667af3993fsi363932eda.20.2024.02.28.08.14.47 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 28 Feb 2024 08:14:48 -0800 (PST) Received-SPF: pass (google.com: domain of linux-kernel+bounces-85334-ouuuleilei=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249; Authentication-Results: mx.google.com; dkim=pass header.i=@intel.com header.s=Intel header.b=HqZ3ndw+; arc=pass (i=1 spf=pass spfdomain=intel.com dkim=pass dkdomain=intel.com dmarc=pass fromdomain=intel.com); spf=pass (google.com: domain of linux-kernel+bounces-85334-ouuuleilei=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) smtp.mailfrom="linux-kernel+bounces-85334-ouuuleilei=gmail.com@vger.kernel.org"; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=intel.com Received: from smtp.subspace.kernel.org (wormhole.subspace.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by am.mirrors.kernel.org (Postfix) with ESMTPS id 515631F23996 for ; Wed, 28 Feb 2024 16:14:28 +0000 (UTC) Received: from localhost.localdomain (localhost.localdomain [127.0.0.1]) by smtp.subspace.kernel.org (Postfix) with ESMTP id AAAD13FB8A; Wed, 28 Feb 2024 16:12:46 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b="HqZ3ndw+" Received: from mgamail.intel.com (mgamail.intel.com [192.198.163.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id D50C415DBC8; Wed, 28 Feb 2024 16:12:39 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=192.198.163.11 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1709136761; cv=none; b=TonsNwD7U/mXRLqltVJcyooRT1wQbGPxifysApBTYfrYYdhivvh11qt8q94qTWDGoRNCd+xnJcGbZiYRywx4ZtvsxMwSLFbIRyMams9eMkHGHaSlVMOCRDfEP5ox5qr/dnBZ+OsW9xtMpRiHndm/H9N+cTWBrEjV/54p9tDHabk= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1709136761; c=relaxed/simple; bh=psRHCo6+Yw3rCVKNpqyRQ2TxSJlnw7iM0+SODvNgFs0=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=heaoHjgYVVOuunViBhcP82K+rMIggepTXiY3YcKuWmENR4cfZc5+lf/b8YnztCsFFhu0QWl9FPcJavJcnHllJ90Q1vdb53F90y8jymrj+78893SZD5SAnbqllxZj+lDTJJ3SmZhxjwyB2xA/izCHjHQcNgtq4yAw+V89IPXw6QY= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=intel.com; spf=pass smtp.mailfrom=intel.com; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b=HqZ3ndw+; arc=none smtp.client-ip=192.198.163.11 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=intel.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=intel.com DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1709136759; x=1740672759; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=psRHCo6+Yw3rCVKNpqyRQ2TxSJlnw7iM0+SODvNgFs0=; b=HqZ3ndw+32w/4LSuMBUb5RUCztZkb84EMYZSoDkmS1VmcqgAZxacjlpl 8stUr+boqRh+HugsnwGflCZHORJZAAmParEGrsaTep+QGJEWzj3cjoxAR he/phEzMT9hwx5frndLUOOy/a4V9xFEiNp5/XaLZWAOYhjNfKG0CTgvaU 3Lcek8WTQq8ugnMNnG12oekRZbF156dZQcmJ6mUJHdoGPuuKsJLZ7vMue Iiq+wzz/NU6jEVb35RKv/VECtzJgh8H7amBPyJuOqkdwovJMshXBqzUuI Nvt6ZIRspaKxjqtc15M0eRgCXN6Y0iwzX8SpmAV87m1J4hwsccupJ9g2X A==; X-IronPort-AV: E=McAfee;i="6600,9927,10998"; a="14185445" X-IronPort-AV: E=Sophos;i="6.06,190,1705392000"; d="scan'208";a="14185445" Received: from orviesa010.jf.intel.com ([10.64.159.150]) by fmvoesa105.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 28 Feb 2024 08:12:37 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.06,190,1705392000"; d="scan'208";a="7418088" Received: from fl31ca102ks0602.deacluster.intel.com (HELO gnr-bkc.deacluster.intel.com) ([10.75.133.163]) by orviesa010.jf.intel.com with ESMTP; 28 Feb 2024 08:12:37 -0800 From: weilin.wang@intel.com To: weilin.wang@intel.com, Ian Rogers , Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Alexander Shishkin , Jiri Olsa , Namhyung Kim , Adrian Hunter , Kan Liang Cc: linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org, Perry Taylor , Samantha Alt , Caleb Biggers Subject: [RFC PATCH v2 1/6] perf stat: Parse and find tpebs events when parsing metrics to prepare for perf record sampling Date: Wed, 28 Feb 2024 11:12:22 -0500 Message-ID: <20240228161227.484886-2-weilin.wang@intel.com> X-Mailer: git-send-email 2.43.0 In-Reply-To: <20240228161227.484886-1-weilin.wang@intel.com> References: <20240228161227.484886-1-weilin.wang@intel.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-getmail-retrieved-from-mailbox: INBOX X-GMAIL-THRID: 1792159921431994610 X-GMAIL-MSGID: 1792159921431994610 From: Weilin Wang Metrics that use tpebs values would use the :retire_latency keyword in formulas. We put all these events into a list and pass the list to perf record to collect their retire latency value. Signed-off-by: Weilin Wang --- tools/perf/builtin-stat.c | 38 +++++++++++++++--- tools/perf/util/metricgroup.c | 73 +++++++++++++++++++++++++++++------ tools/perf/util/metricgroup.h | 10 ++++- tools/perf/util/stat.h | 2 + 4 files changed, 105 insertions(+), 18 deletions(-) diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 6bba1a89d030..5a3093541cff 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -162,6 +162,7 @@ static struct perf_stat_config stat_config = { .ctl_fd = -1, .ctl_fd_ack = -1, .iostat_run = false, + .tpebs_event_size = 0, }; static bool cpus_map_matched(struct evsel *a, struct evsel *b) @@ -686,6 +687,12 @@ static enum counter_recovery stat_handle_error(struct evsel *counter) return COUNTER_FATAL; } +static int __run_perf_record(void) +{ + pr_debug("Prepare perf record for retire_latency\n"); + return 0; +} + static int __run_perf_stat(int argc, const char **argv, int run_idx) { int interval = stat_config.interval; @@ -703,6 +710,15 @@ static int __run_perf_stat(int argc, const char **argv, int run_idx) int err; bool second_pass = false; + //Prepare perf record for sampling event retire_latency before fork and prepare workload + if (stat_config.tpebs_event_size > 0) { + int ret; + + ret = __run_perf_record(); + if (ret) + return ret; + } + if (forks) { if (evlist__prepare_workload(evsel_list, &target, argv, is_pipe, workload_exec_failed_signal) < 0) { perror("failed to prepare workload"); @@ -2106,7 +2122,9 @@ static int add_default_attributes(void) stat_config.metric_no_threshold, stat_config.user_requested_cpu_list, stat_config.system_wide, - &stat_config.metric_events); + &stat_config.metric_events, + &stat_config.tpebs_events, + &stat_config.tpebs_event_size); } if (smi_cost) { @@ -2139,7 +2157,9 @@ static int add_default_attributes(void) stat_config.metric_no_threshold, stat_config.user_requested_cpu_list, stat_config.system_wide, - &stat_config.metric_events); + &stat_config.metric_events, + &stat_config.tpebs_events, + &stat_config.tpebs_event_size); } if (topdown_run) { @@ -2173,7 +2193,9 @@ static int add_default_attributes(void) /*metric_no_threshold=*/true, stat_config.user_requested_cpu_list, stat_config.system_wide, - &stat_config.metric_events) < 0) + &stat_config.metric_events, + &stat_config.tpebs_events, + &stat_config.tpebs_event_size) < 0) return -1; } @@ -2214,7 +2236,9 @@ static int add_default_attributes(void) /*metric_no_threshold=*/true, stat_config.user_requested_cpu_list, stat_config.system_wide, - &stat_config.metric_events) < 0) + &stat_config.metric_events, + /*&stat_config.tpebs_events=*/NULL, + /*stat_config.tpebs_event_size=*/0) < 0) return -1; evlist__for_each_entry(metric_evlist, metric_evsel) { @@ -2736,6 +2760,8 @@ int cmd_stat(int argc, const char **argv) } } + INIT_LIST_HEAD(&stat_config.tpebs_events); + /* * Metric parsing needs to be delayed as metrics may optimize events * knowing the target is system-wide. @@ -2748,7 +2774,9 @@ int cmd_stat(int argc, const char **argv) stat_config.metric_no_threshold, stat_config.user_requested_cpu_list, stat_config.system_wide, - &stat_config.metric_events); + &stat_config.metric_events, + &stat_config.tpebs_events, + &stat_config.tpebs_event_size); zfree(&metrics); if (ret) { diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c index b24a1c177a80..7a360a9f5b72 100644 --- a/tools/perf/util/metricgroup.c +++ b/tools/perf/util/metricgroup.c @@ -277,7 +277,8 @@ static bool contains_metric_id(struct evsel **metric_events, int num_events, */ static int setup_metric_events(const char *pmu, struct hashmap *ids, struct evlist *metric_evlist, - struct evsel ***out_metric_events) + struct evsel ***out_metric_events, + size_t tpebs_event_size) { struct evsel **metric_events; const char *metric_id; @@ -286,7 +287,7 @@ static int setup_metric_events(const char *pmu, struct hashmap *ids, bool all_pmus = !strcmp(pmu, "all") || perf_pmus__num_core_pmus() == 1 || !is_pmu_core(pmu); *out_metric_events = NULL; - ids_size = hashmap__size(ids); + ids_size = hashmap__size(ids) - tpebs_event_size; metric_events = calloc(ids_size + 1, sizeof(void *)); if (!metric_events) @@ -323,6 +324,7 @@ static int setup_metric_events(const char *pmu, struct hashmap *ids, } } if (matched_events < ids_size) { + pr_debug("Error: matched_events = %lu, ids_size = %lu\n", matched_events, ids_size); free(metric_events); return -EINVAL; } @@ -670,7 +672,9 @@ static int decode_all_metric_ids(struct evlist *perf_evlist, const char *modifie static int metricgroup__build_event_string(struct strbuf *events, const struct expr_parse_ctx *ctx, const char *modifier, - bool group_events) + bool group_events, + struct list_head *tpebs_events __maybe_unused, + size_t *tpebs_event_size) { struct hashmap_entry *cur; size_t bkt; @@ -683,9 +687,42 @@ static int metricgroup__build_event_string(struct strbuf *events, hashmap__for_each_entry(ctx->ids, cur, bkt) { const char *sep, *rsep, *id = cur->pkey; enum perf_tool_event ev; + char *p = strstr(id, ":retire_latency"); pr_debug("found event %s\n", id); + if (p) { + struct tpebs_event *new_event = malloc(sizeof(struct tpebs_event)); + char *name; + char *at; + + pr_debug("event name %s\n", id); + new_event->tpebs_name = strdup(id); + *p = '\0'; + name = malloc(strlen(id) + 2); + if (!name) + return -ENOMEM; + + at = strchr(id, '@'); + if (at != NULL) { + pr_debug("tpebs name %s\n", id); + *at = '/'; + at = strchr(id, '@'); + *at = '/'; + strcpy(name, id); + strcat(name, "p"); + } else { + strcpy(name, id); + strcat(name, ":p"); + } + new_event->name = name; + *tpebs_event_size += 1; + pr_debug("retire_latency required, tpebs_event_size=%lu, new_event=%s\n", + *tpebs_event_size, new_event->name); + list_add_tail(&new_event->nd, tpebs_events); + continue; + } + /* Always move tool events outside of the group. */ ev = perf_tool_event__from_str(id); if (ev != PERF_TOOL_NONE) { @@ -1449,7 +1486,8 @@ static int build_combined_expr_ctx(const struct list_head *metric_list, static int parse_ids(bool metric_no_merge, struct perf_pmu *fake_pmu, struct expr_parse_ctx *ids, const char *modifier, bool group_events, const bool tool_events[PERF_TOOL_MAX], - struct evlist **out_evlist) + struct evlist **out_evlist, struct list_head *tpebs_events, + size_t *tpebs_event_size) { struct parse_events_error parse_error; struct evlist *parsed_evlist; @@ -1492,7 +1530,7 @@ static int parse_ids(bool metric_no_merge, struct perf_pmu *fake_pmu, } } ret = metricgroup__build_event_string(&events, ids, modifier, - group_events); + group_events, tpebs_events, tpebs_event_size); if (ret) return ret; @@ -1531,7 +1569,9 @@ static int parse_groups(struct evlist *perf_evlist, bool system_wide, struct perf_pmu *fake_pmu, struct rblist *metric_events_list, - const struct pmu_metrics_table *table) + const struct pmu_metrics_table *table, + struct list_head *tpebs_events, + size_t *tpebs_event_size) { struct evlist *combined_evlist = NULL; LIST_HEAD(metric_list); @@ -1563,7 +1603,8 @@ static int parse_groups(struct evlist *perf_evlist, /*modifier=*/NULL, /*group_events=*/false, tool_events, - &combined_evlist); + &combined_evlist, + tpebs_events, tpebs_event_size); } if (combined) expr__ctx_free(combined); @@ -1618,14 +1659,15 @@ static int parse_groups(struct evlist *perf_evlist, } if (!metric_evlist) { ret = parse_ids(metric_no_merge, fake_pmu, m->pctx, m->modifier, - m->group_events, tool_events, &m->evlist); + m->group_events, tool_events, &m->evlist, + tpebs_events, tpebs_event_size); if (ret) goto out; metric_evlist = m->evlist; } ret = setup_metric_events(fake_pmu ? "all" : m->pmu, m->pctx->ids, - metric_evlist, &metric_events); + metric_evlist, &metric_events, *tpebs_event_size); if (ret) { pr_err("Cannot resolve IDs for %s: %s\n", m->metric_name, m->metric_expr); @@ -1692,16 +1734,21 @@ int metricgroup__parse_groups(struct evlist *perf_evlist, bool metric_no_threshold, const char *user_requested_cpu_list, bool system_wide, - struct rblist *metric_events) + struct rblist *metric_events, + struct list_head *tpebs_events, + size_t *tpebs_event_size) { const struct pmu_metrics_table *table = pmu_metrics_table__find(); + pr_debug("Test debugging\n"); + if (!table) return -EINVAL; return parse_groups(perf_evlist, pmu, str, metric_no_group, metric_no_merge, metric_no_threshold, user_requested_cpu_list, system_wide, - /*fake_pmu=*/NULL, metric_events, table); + /*fake_pmu=*/NULL, metric_events, table, tpebs_events, + tpebs_event_size); } int metricgroup__parse_groups_test(struct evlist *evlist, @@ -1715,7 +1762,9 @@ int metricgroup__parse_groups_test(struct evlist *evlist, /*metric_no_threshold=*/false, /*user_requested_cpu_list=*/NULL, /*system_wide=*/false, - &perf_pmu__fake, metric_events, table); + &perf_pmu__fake, metric_events, table, + /*tpebs_events=*/NULL, + /*tpebs_event_size=*/0); } struct metricgroup__has_metric_data { diff --git a/tools/perf/util/metricgroup.h b/tools/perf/util/metricgroup.h index d5325c6ec8e1..7c24ed768ff3 100644 --- a/tools/perf/util/metricgroup.h +++ b/tools/perf/util/metricgroup.h @@ -66,6 +66,12 @@ struct metric_expr { int runtime; }; +struct tpebs_event { + struct list_head nd; + const char *name; + const char *tpebs_name; +}; + struct metric_event *metricgroup__lookup(struct rblist *metric_events, struct evsel *evsel, bool create); @@ -77,7 +83,9 @@ int metricgroup__parse_groups(struct evlist *perf_evlist, bool metric_no_threshold, const char *user_requested_cpu_list, bool system_wide, - struct rblist *metric_events); + struct rblist *metric_events, + struct list_head *tpebs_events, + size_t *tpebs_event_size); int metricgroup__parse_groups_test(struct evlist *evlist, const struct pmu_metrics_table *table, const char *str, diff --git a/tools/perf/util/stat.h b/tools/perf/util/stat.h index d6e5c8787ba2..b987960df3c5 100644 --- a/tools/perf/util/stat.h +++ b/tools/perf/util/stat.h @@ -109,6 +109,8 @@ struct perf_stat_config { struct cpu_aggr_map *cpus_aggr_map; u64 *walltime_run; struct rblist metric_events; + struct list_head tpebs_events; + size_t tpebs_event_size; int ctl_fd; int ctl_fd_ack; bool ctl_fd_close;