From nobody Fri Dec 19 20:54:09 2025 Received: from mail-yb1-f202.google.com (mail-yb1-f202.google.com [209.85.219.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id A87EF1A706F for ; Wed, 8 Jan 2025 05:34:53 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.219.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1736314495; cv=none; b=Hm98I+4GUWBLxta/+C6VETFYQnkx+/Aa7MA4/clpt11e9CjKYE8u85uxYTkvURjqvCWQ2h9mZbDJbSjMFH/7kKJqkzaIytW6kWlqfCmyMcFjN3DPozCPpvGr1HXDeRxCoqUDW/oKbMGU+/Kv+Z9kTr/O26Q62QwKUznHkRH2Fsk= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1736314495; c=relaxed/simple; bh=ywjvsWJNxLmzM3YIwbYYswpNEHZxKvufAxgc8RsNSbc=; h=Date:In-Reply-To:Message-Id:Mime-Version:References:Subject:From: To:Content-Type; b=jpnQOXWppVYaUKtMopM81y4hFZFTuCQevhIXwC+ENX1LAbxFcJsF7GPz3WsZEXtRL9livlTjA6aeEfT0/6ddgOEvBVagOuMQrXyBQpASC9aEjganBe0RLbM3KeQaAcOi1i4BCkuIc0cVwVddJIy0HMl2yxA5Kw7crkkGuKiisII= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=gHShGJIV; arc=none smtp.client-ip=209.85.219.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="gHShGJIV" Received: by mail-yb1-f202.google.com with SMTP id 3f1490d57ef6-e544c8dae18so20213390276.1 for ; Tue, 07 Jan 2025 21:34:53 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1736314493; x=1736919293; darn=vger.kernel.org; h=to:from:subject:references:mime-version:message-id:in-reply-to:date :from:to:cc:subject:date:message-id:reply-to; bh=XcjD+oXNLY7oywMEC3LvBmYCeG/3kMVQ1vWw9hGSJM0=; b=gHShGJIVO0PMSPbryqkCyPOs2pWEZizw06AbyAsV02oiZngRSaMVq0XpPi54TjJiUE xDEC2TLIj2h9j3ZF5z3ljBYbp11AGLN4yrpoaNaREfetB4dntcKuVHmmeOufsomR9yeh K+lWCSt12rKR8qUeR82FS6TpBd4RAIBFJr64nbr6clu5vfTGr/B1/BMzvG3R1Tn30gV+ dmFCw+/ntc8xJiw18nSbvys1+Ye9G6SoyctCUVVI+ERZKsrDcMevnCtTEtlGxA0IGC+3 5PvcBi4oBOXJP9zkQ4HZmHi6x0ZnLRoqILqFD92p2WQ03lOVcG5PrRP9Rd6bMJhppfPR F++g== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1736314493; x=1736919293; h=to:from:subject:references:mime-version:message-id:in-reply-to:date :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=XcjD+oXNLY7oywMEC3LvBmYCeG/3kMVQ1vWw9hGSJM0=; b=PzJdsrLFIQvRuVKoImYo2JwmCEV5KoYzzdXqolXBQ57TqrFfKl5jOlnb/+jRDv60pN qXpjPoKpCYyuvhppLja57TpV7wJi6rb6CChlabMadTtb062zJMVO/Iu+RnpVb2V47pF/ SAUHYRyys9iZByv827UvdQiN/epHEd+xR3bph9xfhPzUQioxzNHH/lU6J0fdaCxRPYqC hrgVg+6bvPVqY3E73QluzaxQG80+S3oNKH00xvEG0U/Hpf01e0QxHVV6IAsv5gNL53wL MPnDxCFqIB/OtKu1xEwpmbSscONg97UfwfQmx41DR7UYCHVOKEJAKzJZDe3Qyy0OwFgy Xl5A== X-Forwarded-Encrypted: i=1; AJvYcCUnk3rhLzw48mjdqyBh7404enNsw1824AFuJGiiivwSxVFQdD04VFLB1Npj+KScG5x5oa64n+2YkYhJoOI=@vger.kernel.org X-Gm-Message-State: AOJu0YwCRjTWy+ZxDAHmpybiRRM8M8msBz5HJUmzLH7KSMUd/ZRYWy8J H7PfNxtqOJn525tWehQklAewzz3kBe2uh/jks4HHCt4WezJcy3XOEIxLSZqXElenlxAyp5rhwsz Cwtl5LA== X-Google-Smtp-Source: AGHT+IEW9+WqALHzqTdZt6mmp40Bu6OzzY4Eg/OXgZx1NeC6753KkNiIRxAOLHlol1lIM4PvLR54hHNmI92D X-Received: from irogers.svl.corp.google.com ([2620:15c:2c5:11:7ae:a887:5504:e3ca]) (user=irogers job=sendgmr) by 2002:a5b:642:0:b0:e39:6fdc:5567 with SMTP id 3f1490d57ef6-e54eddd285bmr4357276.0.1736314492910; Tue, 07 Jan 2025 21:34:52 -0800 (PST) Date: Tue, 7 Jan 2025 21:34:28 -0800 In-Reply-To: <20250108053428.1724490-1-irogers@google.com> Message-Id: <20250108053428.1724490-5-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250108053428.1724490-1-irogers@google.com> X-Mailer: git-send-email 2.47.1.613.gc27f4b7a9f-goog Subject: [PATCH v4 4/4] perf parse-events: Add "cpu" term to set the CPU an event is recorded on From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , Leo Yan , Yoshihiro Furudera , Weilin Wang , Andi Kleen , James Clark , Dominique Martinet , Yicong Yang , linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" The -C option allows the CPUs for a list of events to be specified but its not possible to set the CPU for a single event. Add a term to allow this. The term isn't a general CPU list due to ',' already being a special character in event parsing instead multiple cpu=3D terms may be provided and they will be merged/unioned together. An example of mixing different types of events counted on different CPUs: ``` $ perf stat -A -C 0,4-5,8 -e "instructions/cpu=3D0/,l1d-misses/cpu=3D4,cpu= =3D5/,inst_retired.any/cpu=3D8/,cycles" -a sleep 0.1 Performance counter stats for 'system wide': CPU0 6,979,225 instructions/cpu=3D0/ # 0.89= insn per cycle CPU4 75,138 cpu/l1d-misses/ CPU5 1,418,939 cpu/l1d-misses/ CPU8 797,553 cpu/inst_retired.any,cpu=3D8/ CPU0 7,845,302 cycles CPU4 6,546,859 cycles CPU5 185,915,438 cycles CPU8 2,065,668 cycles 0.112449242 seconds time elapsed ``` Signed-off-by: Ian Rogers --- tools/perf/Documentation/perf-list.txt | 9 +++ tools/perf/util/evsel_config.h | 1 + tools/perf/util/parse-events.c | 76 +++++++++++++++++++++----- tools/perf/util/parse-events.h | 3 +- tools/perf/util/parse-events.l | 1 + tools/perf/util/pmu.c | 3 +- 6 files changed, 76 insertions(+), 17 deletions(-) diff --git a/tools/perf/Documentation/perf-list.txt b/tools/perf/Documentat= ion/perf-list.txt index d0c65fad419a..d98de9adc515 100644 --- a/tools/perf/Documentation/perf-list.txt +++ b/tools/perf/Documentation/perf-list.txt @@ -289,6 +289,15 @@ Sums up the event counts for all hardware threads in a= core, e.g.: =20 perf stat -e cpu/event=3D0,umask=3D0x3,percore=3D1/ =20 +cpu: + +Specifies the CPU to open the event upon. The value may be repeated to +specify opening the event on multiple CPUs: + + + perf stat -e instructions/cpu=3D0,cpu=3D2/,cycles/cpu=3D1,cpu=3D2/ -a sl= eep 1 + perf stat -e data_read/cpu=3D0/,data_write/cpu=3D1/ -a sleep 1 + =20 EVENT GROUPS ------------ diff --git a/tools/perf/util/evsel_config.h b/tools/perf/util/evsel_config.h index af52a1516d0b..94a1e9cf73d6 100644 --- a/tools/perf/util/evsel_config.h +++ b/tools/perf/util/evsel_config.h @@ -48,6 +48,7 @@ struct evsel_config_term { u32 aux_sample_size; u64 cfg_chg; char *str; + int cpu; } val; bool weak; }; diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 4bda4141e9e7..ec8429f0cf6d 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -7,6 +7,7 @@ #include #include #include +#include "cpumap.h" #include "term.h" #include "env.h" #include "evlist.h" @@ -178,6 +179,26 @@ static char *get_config_name(const struct parse_events= _terms *head_terms) return get_config_str(head_terms, PARSE_EVENTS__TERM_TYPE_NAME); } =20 +static struct perf_cpu_map *get_config_cpu(const struct parse_events_terms= *head_terms) +{ + struct parse_events_term *term; + struct perf_cpu_map *cpus =3D NULL; + + if (!head_terms) + return NULL; + + list_for_each_entry(term, &head_terms->terms, list) { + if (term->type_term =3D=3D PARSE_EVENTS__TERM_TYPE_CPU) { + struct perf_cpu_map *cpu =3D perf_cpu_map__new_int(term->val.num); + + perf_cpu_map__merge(&cpus, cpu); + perf_cpu_map__put(cpu); + } + } + + return cpus; +} + /** * fix_raw - For each raw term see if there is an event (aka alias) in pmu= that * matches the raw's string value. If the string value matches an @@ -441,11 +462,12 @@ int parse_events_add_cache(struct list_head *list, in= t *idx, const char *name, bool found_supported =3D false; const char *config_name =3D get_config_name(parsed_terms); const char *metric_id =3D get_config_metric_id(parsed_terms); + struct perf_cpu_map *cpus =3D get_config_cpu(parsed_terms); + int ret =3D 0; =20 while ((pmu =3D perf_pmus__scan(pmu)) !=3D NULL) { LIST_HEAD(config_terms); struct perf_event_attr attr; - int ret; =20 if (parse_events__filter_pmu(parse_state, pmu)) continue; @@ -460,7 +482,7 @@ int parse_events_add_cache(struct list_head *list, int = *idx, const char *name, perf_pmu__auto_merge_stats(pmu), /*alternate_hw_config=3D*/PERF_COUNT_HW_MAX); if (ret) - return ret; + goto out_err; continue; } =20 @@ -480,21 +502,27 @@ int parse_events_add_cache(struct list_head *list, in= t *idx, const char *name, =20 if (parsed_terms) { if (config_attr(&attr, parsed_terms, parse_state->error, - config_term_common)) - return -EINVAL; - - if (get_config_terms(parsed_terms, &config_terms)) - return -ENOMEM; + config_term_common)) { + ret =3D -EINVAL; + goto out_err; + } + if (get_config_terms(parsed_terms, &config_terms)) { + ret =3D -ENOMEM; + goto out_err; + } } =20 if (__add_event(list, idx, &attr, /*init_attr*/true, config_name ?: name, metric_id, pmu, &config_terms, /*auto_merge_stats=3D*/false, - /*cpu_list=3D*/NULL, - /*alternate_hw_config=3D*/PERF_COUNT_HW_MAX) =3D=3D NULL) - return -ENOMEM; + cpus, /*alternate_hw_config=3D*/PERF_COUNT_HW_MAX) =3D=3D NULL) + ret =3D -ENOMEM; =20 free_config_terms(&config_terms); + if (ret) + goto out_err; } +out_err: + perf_cpu_map__put(cpus); return found_supported ? 0 : -EINVAL; } =20 @@ -808,6 +836,7 @@ const char *parse_events__term_type_str(enum parse_even= ts__term_type term_type) [PARSE_EVENTS__TERM_TYPE_RAW] =3D "raw", [PARSE_EVENTS__TERM_TYPE_LEGACY_CACHE] =3D "legacy-cache", [PARSE_EVENTS__TERM_TYPE_HARDWARE] =3D "hardware", + [PARSE_EVENTS__TERM_TYPE_CPU] =3D "cpu", }; if ((unsigned int)term_type >=3D __PARSE_EVENTS__TERM_TYPE_NR) return "unknown term"; @@ -837,6 +866,7 @@ config_term_avail(enum parse_events__term_type term_typ= e, struct parse_events_er case PARSE_EVENTS__TERM_TYPE_METRIC_ID: case PARSE_EVENTS__TERM_TYPE_SAMPLE_PERIOD: case PARSE_EVENTS__TERM_TYPE_PERCORE: + case PARSE_EVENTS__TERM_TYPE_CPU: return true; case PARSE_EVENTS__TERM_TYPE_USER: case PARSE_EVENTS__TERM_TYPE_SAMPLE_FREQ: @@ -984,6 +1014,15 @@ do { \ return -EINVAL; } break; + case PARSE_EVENTS__TERM_TYPE_CPU: + CHECK_TYPE_VAL(NUM); + if (term->val.num >=3D (u64)cpu__max_present_cpu().cpu) { + parse_events_error__handle(err, term->err_val, + strdup("too big"), + NULL); + return -EINVAL; + } + break; case PARSE_EVENTS__TERM_TYPE_DRV_CFG: case PARSE_EVENTS__TERM_TYPE_USER: case PARSE_EVENTS__TERM_TYPE_LEGACY_CACHE: @@ -1111,6 +1150,7 @@ static int config_term_tracepoint(struct perf_event_a= ttr *attr, case PARSE_EVENTS__TERM_TYPE_RAW: case PARSE_EVENTS__TERM_TYPE_LEGACY_CACHE: case PARSE_EVENTS__TERM_TYPE_HARDWARE: + case PARSE_EVENTS__TERM_TYPE_CPU: default: if (err) { parse_events_error__handle(err, term->err_term, @@ -1245,6 +1285,7 @@ do { \ case PARSE_EVENTS__TERM_TYPE_RAW: case PARSE_EVENTS__TERM_TYPE_LEGACY_CACHE: case PARSE_EVENTS__TERM_TYPE_HARDWARE: + case PARSE_EVENTS__TERM_TYPE_CPU: default: break; } @@ -1299,6 +1340,7 @@ static int get_config_chgs(struct perf_pmu *pmu, stru= ct parse_events_terms *head case PARSE_EVENTS__TERM_TYPE_RAW: case PARSE_EVENTS__TERM_TYPE_LEGACY_CACHE: case PARSE_EVENTS__TERM_TYPE_HARDWARE: + case PARSE_EVENTS__TERM_TYPE_CPU: default: break; } @@ -1343,6 +1385,7 @@ static int __parse_events_add_numeric(struct parse_ev= ents_state *parse_state, struct perf_event_attr attr; LIST_HEAD(config_terms); const char *name, *metric_id; + struct perf_cpu_map *cpus; int ret; =20 memset(&attr, 0, sizeof(attr)); @@ -1364,10 +1407,11 @@ static int __parse_events_add_numeric(struct parse_= events_state *parse_state, =20 name =3D get_config_name(head_config); metric_id =3D get_config_metric_id(head_config); + cpus =3D get_config_cpu(head_config); ret =3D __add_event(list, &parse_state->idx, &attr, /*init_attr*/true, na= me, - metric_id, pmu, &config_terms, /*auto_merge_stats=3D*/false, - /*cpu_list=3D*/NULL, /*alternate_hw_config=3D*/PERF_COUNT_HW_MAX - ) =3D=3D NULL ? -ENOMEM : 0; + metric_id, pmu, &config_terms, /*auto_merge_stats=3D*/false, + cpus, /*alternate_hw_config=3D*/PERF_COUNT_HW_MAX) ? 0 : -ENOMEM; + perf_cpu_map__put(cpus); free_config_terms(&config_terms); return ret; } @@ -1427,6 +1471,7 @@ static int parse_events_add_pmu(struct parse_events_s= tate *parse_state, LIST_HEAD(config_terms); struct parse_events_terms parsed_terms; bool alias_rewrote_terms =3D false; + struct perf_cpu_map *term_cpu =3D NULL; =20 if (verbose > 1) { struct strbuf sb; @@ -1521,11 +1566,12 @@ static int parse_events_add_pmu(struct parse_events= _state *parse_state, return -EINVAL; } =20 + term_cpu =3D get_config_cpu(&parsed_terms); evsel =3D __add_event(list, &parse_state->idx, &attr, /*init_attr=3D*/tru= e, get_config_name(&parsed_terms), get_config_metric_id(&parsed_terms), pmu, - &config_terms, auto_merge_stats, /*cpu_list=3D*/NULL, - alternate_hw_config); + &config_terms, auto_merge_stats, term_cpu, alternate_hw_config); + perf_cpu_map__put(term_cpu); if (!evsel) { parse_events_terms__exit(&parsed_terms); return -ENOMEM; diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index e176a34ab088..ab242f671031 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -80,7 +80,8 @@ enum parse_events__term_type { PARSE_EVENTS__TERM_TYPE_RAW, PARSE_EVENTS__TERM_TYPE_LEGACY_CACHE, PARSE_EVENTS__TERM_TYPE_HARDWARE, -#define __PARSE_EVENTS__TERM_TYPE_NR (PARSE_EVENTS__TERM_TYPE_HARDWARE + 1) + PARSE_EVENTS__TERM_TYPE_CPU, +#define __PARSE_EVENTS__TERM_TYPE_NR (PARSE_EVENTS__TERM_TYPE_CPU + 1) }; =20 struct parse_events_term { diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l index bf7f73548605..14bff6f9731c 100644 --- a/tools/perf/util/parse-events.l +++ b/tools/perf/util/parse-events.l @@ -324,6 +324,7 @@ aux-output { return term(yyscanner, PARSE_EVENTS__TERM= _TYPE_AUX_OUTPUT); } aux-action { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_AUX_ACTION); } aux-sample-size { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_AUX_SAMP= LE_SIZE); } metric-id { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_METRIC_ID); } +cpu { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_CPU); } cpu-cycles|cycles { return hw_term(yyscanner, PERF_COUNT_HW_CPU_CYCLES)= ; } stalled-cycles-frontend|idle-cycles-frontend { return hw_term(yyscanner, P= ERF_COUNT_HW_STALLED_CYCLES_FRONTEND); } stalled-cycles-backend|idle-cycles-backend { return hw_term(yyscanner, PER= F_COUNT_HW_STALLED_CYCLES_BACKEND); } diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index 6206c8fe2bf9..cffb4eb2696b 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -1400,7 +1400,7 @@ static int pmu_config_term(const struct perf_pmu *pmu, break; case PARSE_EVENTS__TERM_TYPE_USER: /* Not hardcoded. */ return -EINVAL; - case PARSE_EVENTS__TERM_TYPE_NAME ... PARSE_EVENTS__TERM_TYPE_HARDWARE: + case PARSE_EVENTS__TERM_TYPE_NAME ... PARSE_EVENTS__TERM_TYPE_CPU: /* Skip non-config terms. */ break; default: @@ -1775,6 +1775,7 @@ int perf_pmu__for_each_format(struct perf_pmu *pmu, v= oid *state, pmu_format_call "aux-output", "aux-action=3D(pause|resume|start-paused)", "aux-sample-size=3Dnumber", + "cpu=3Dnumber", }; struct perf_pmu_format *format; int ret; --=20 2.47.1.613.gc27f4b7a9f-goog