From nobody Thu Jan 1 09:12:23 2026 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id AA76AC25B47 for ; Tue, 24 Oct 2023 22:25:51 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1344440AbjJXWZu (ORCPT ); Tue, 24 Oct 2023 18:25:50 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:38988 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S234965AbjJXWZ1 (ORCPT ); Tue, 24 Oct 2023 18:25:27 -0400 Received: from mail-yb1-xb49.google.com (mail-yb1-xb49.google.com [IPv6:2607:f8b0:4864:20::b49]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 6C2BA1FC1 for ; Tue, 24 Oct 2023 15:24:51 -0700 (PDT) Received: by mail-yb1-xb49.google.com with SMTP id 3f1490d57ef6-d8486b5e780so5794977276.0 for ; Tue, 24 Oct 2023 15:24:51 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1698186284; x=1698791084; 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=mFUu7p9S5YeiUqPkIKu4IgMBXIymBeuUApOrPlH1y/o=; b=X9yI6FUy1ePkOXhvKAD9GqV+iHt4OfjBw2bHjlDjqYkkOzOO8SJo0aitaFi88R9q4d PcsXtoac0P1kCi+WPk/c/B3aRHDMGdDS8BJ0Po4dUsLXYoUQLMUqo+RAdJb7GtxmbNT9 1NHVgKxorsA+cswmelXKymelKOdD9jHsZIsOZJRqRZVGUY1nR9vfSbySycHz8irrqpLW GSiU4yVbpMRrQA8NZ1Jr0uOVSkSt8hd3hNpKvKymQMc7NcnuWXOJ0mSFrP/spgzB89fi 1gOCBSa3IQRmBuNnA6JkxpIpm7gccogElMpep9X92n9LAAdy7MKeeAPkD0eLHmJVzB3z SsSg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1698186284; x=1698791084; 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=mFUu7p9S5YeiUqPkIKu4IgMBXIymBeuUApOrPlH1y/o=; b=ZC5ICiOxTYJgNtoiCdn8eItYeKYG2KatodKKo9XYD1Yd+BA3PO4PET/jI0e012Fmbu yydFPkQFGJj/WSZG0VnJR4R0mliu6Us9XTXQuqpc2qlstRP4QhA0Rs0PktM+iZQctvqV LoV/9YMua/XcoapRrhqEWvdUFg7GF+TVOrN61Dyy2VU3fbDvKCpZqtM1d+EQjELYY/z5 kuoan+YyWTnPkUIOeagGDT0FevzNTIymbRmMjPy3oOoCOu1Hn/ONTGJ/6ETPknlgopdr Leos2o6DDr+2PBdgB73ujV8k6H1L9bunKyokpOrSjn/NLyKTXppMOhm1Y05F+2jzLmEB Ew+w== X-Gm-Message-State: AOJu0Yx4EOe5KknvriiR6Dhxnk97o8OQ9d9RHczTN+vutwTiNdKiCrZJ UgoPAytPsafMt1X1Ru60fnW5KY2kc6j/ X-Google-Smtp-Source: AGHT+IGSu0bx18CPVFKMs58nB3xLIn9DFr9Q0XU1S64CM6ktTDhMVgupYp1iv/ErkMArmokrPupa/XSu0TnC X-Received: from irogers.svl.corp.google.com ([2620:15c:2a3:200:93d2:18cc:4d63:45ba]) (user=irogers job=sendgmr) by 2002:a25:324b:0:b0:d9b:f88c:ca2b with SMTP id y72-20020a25324b000000b00d9bf88cca2bmr258286yby.11.1698186284456; Tue, 24 Oct 2023 15:24:44 -0700 (PDT) Date: Tue, 24 Oct 2023 15:23:17 -0700 In-Reply-To: <20231024222353.3024098-1-irogers@google.com> Message-Id: <20231024222353.3024098-15-irogers@google.com> Mime-Version: 1.0 References: <20231024222353.3024098-1-irogers@google.com> X-Mailer: git-send-email 2.42.0.758.gaed0368e0e-goog Subject: [PATCH v3 14/50] perf mmap: Lazily initialize zstd streams From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Mark Rutland , Alexander Shishkin , Jiri Olsa , Namhyung Kim , Ian Rogers , Adrian Hunter , Nick Terrell , Kan Liang , Andi Kleen , Leo Yan , Song Liu , Sandipan Das , James Clark , Anshuman Khandual , Miguel Ojeda , Liam Howlett , Yang Jihong , Athira Rajeev , Kajol Jain , K Prateek Nayak , Sean Christopherson , Yanteng Si , Ravi Bangoria , German Gomez , Changbin Du , Paolo Bonzini , Masami Hiramatsu , liuwenyu , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Zstd streams create dictionaries that can require significant RAM, especially when there is one per-CPU. Tools like perf record won't use the streams without the -z option, and so the creation of the streams is pure overhead. Switch to creating the streams on first use. Signed-off-by: Ian Rogers --- tools/perf/builtin-record.c | 26 ++++++++++----- tools/perf/util/compress.h | 6 ++-- tools/perf/util/mmap.c | 5 ++- tools/perf/util/mmap.h | 1 - tools/perf/util/zstd.c | 63 +++++++++++++++++++------------------ 5 files changed, 58 insertions(+), 43 deletions(-) diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 8ec818568662..9b4f3805ca92 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -270,7 +270,7 @@ static int record__write(struct record *rec, struct mma= p *map __maybe_unused, =20 static int record__aio_enabled(struct record *rec); static int record__comp_enabled(struct record *rec); -static size_t zstd_compress(struct perf_session *session, struct mmap *map, +static ssize_t zstd_compress(struct perf_session *session, struct mmap *ma= p, void *dst, size_t dst_size, void *src, size_t src_size); =20 #ifdef HAVE_AIO_SUPPORT @@ -405,9 +405,13 @@ static int record__aio_pushfn(struct mmap *map, void *= to, void *buf, size_t size */ =20 if (record__comp_enabled(aio->rec)) { - size =3D zstd_compress(aio->rec->session, NULL, aio->data + aio->size, - mmap__mmap_len(map) - aio->size, - buf, size); + ssize_t compressed =3D zstd_compress(aio->rec->session, NULL, aio->data = + aio->size, + mmap__mmap_len(map) - aio->size, + buf, size); + if (compressed < 0) + return (int)compressed; + + size =3D compressed; } else { memcpy(aio->data + aio->size, buf, size); } @@ -633,7 +637,13 @@ static int record__pushfn(struct mmap *map, void *to, = void *bf, size_t size) struct record *rec =3D to; =20 if (record__comp_enabled(rec)) { - size =3D zstd_compress(rec->session, map, map->data, mmap__mmap_len(map)= , bf, size); + ssize_t compressed =3D zstd_compress(rec->session, map, map->data, + mmap__mmap_len(map), bf, size); + + if (compressed < 0) + return (int)compressed; + + size =3D compressed; bf =3D map->data; } =20 @@ -1527,10 +1537,10 @@ static size_t process_comp_header(void *record, siz= e_t increment) return size; } =20 -static size_t zstd_compress(struct perf_session *session, struct mmap *map, +static ssize_t zstd_compress(struct perf_session *session, struct mmap *ma= p, void *dst, size_t dst_size, void *src, size_t src_size) { - size_t compressed; + ssize_t compressed; size_t max_record_size =3D PERF_SAMPLE_MAX_SIZE - sizeof(struct perf_reco= rd_compressed) - 1; struct zstd_data *zstd_data =3D &session->zstd_data; =20 @@ -1539,6 +1549,8 @@ static size_t zstd_compress(struct perf_session *sess= ion, struct mmap *map, =20 compressed =3D zstd_compress_stream_to_records(zstd_data, dst, dst_size, = src, src_size, max_record_size, process_comp_header); + if (compressed < 0) + return compressed; =20 if (map && map->file) { thread->bytes_transferred +=3D src_size; diff --git a/tools/perf/util/compress.h b/tools/perf/util/compress.h index 0cd3369af2a4..9eb6eb5bf038 100644 --- a/tools/perf/util/compress.h +++ b/tools/perf/util/compress.h @@ -3,6 +3,7 @@ #define PERF_COMPRESS_H =20 #include +#include #ifdef HAVE_ZSTD_SUPPORT #include #endif @@ -21,6 +22,7 @@ struct zstd_data { #ifdef HAVE_ZSTD_SUPPORT ZSTD_CStream *cstream; ZSTD_DStream *dstream; + int comp_level; #endif }; =20 @@ -29,7 +31,7 @@ struct zstd_data { int zstd_init(struct zstd_data *data, int level); int zstd_fini(struct zstd_data *data); =20 -size_t zstd_compress_stream_to_records(struct zstd_data *data, void *dst, = size_t dst_size, +ssize_t zstd_compress_stream_to_records(struct zstd_data *data, void *dst,= size_t dst_size, void *src, size_t src_size, size_t max_record_size, size_t process_header(void *record, size_t increment)); =20 @@ -48,7 +50,7 @@ static inline int zstd_fini(struct zstd_data *data __mayb= e_unused) } =20 static inline -size_t zstd_compress_stream_to_records(struct zstd_data *data __maybe_unus= ed, +ssize_t zstd_compress_stream_to_records(struct zstd_data *data __maybe_unu= sed, void *dst __maybe_unused, size_t dst_size __maybe_unused, void *src __maybe_unused, size_t src_size __maybe_unused, size_t max_record_size __maybe_unused, diff --git a/tools/perf/util/mmap.c b/tools/perf/util/mmap.c index 49093b21ee2d..122ee198a86e 100644 --- a/tools/perf/util/mmap.c +++ b/tools/perf/util/mmap.c @@ -295,15 +295,14 @@ int mmap__mmap(struct mmap *map, struct mmap_params *= mp, int fd, struct perf_cpu =20 map->core.flush =3D mp->flush; =20 - map->comp_level =3D mp->comp_level; #ifndef PYTHON_PERF - if (zstd_init(&map->zstd_data, map->comp_level)) { + if (zstd_init(&map->zstd_data, mp->comp_level)) { pr_debug2("failed to init mmap compressor, error %d\n", errno); return -1; } #endif =20 - if (map->comp_level && !perf_mmap__aio_enabled(map)) { + if (mp->comp_level && !perf_mmap__aio_enabled(map)) { map->data =3D mmap(NULL, mmap__mmap_len(map), PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, 0, 0); if (map->data =3D=3D MAP_FAILED) { diff --git a/tools/perf/util/mmap.h b/tools/perf/util/mmap.h index f944c3cd5efa..0df6e1621c7e 100644 --- a/tools/perf/util/mmap.h +++ b/tools/perf/util/mmap.h @@ -39,7 +39,6 @@ struct mmap { #endif struct mmap_cpu_mask affinity_mask; void *data; - int comp_level; struct perf_data_file *file; struct zstd_data zstd_data; }; diff --git a/tools/perf/util/zstd.c b/tools/perf/util/zstd.c index 48dd2b018c47..57027e0ac7b6 100644 --- a/tools/perf/util/zstd.c +++ b/tools/perf/util/zstd.c @@ -7,35 +7,9 @@ =20 int zstd_init(struct zstd_data *data, int level) { - size_t ret; - - data->dstream =3D ZSTD_createDStream(); - if (data->dstream =3D=3D NULL) { - pr_err("Couldn't create decompression stream.\n"); - return -1; - } - - ret =3D ZSTD_initDStream(data->dstream); - if (ZSTD_isError(ret)) { - pr_err("Failed to initialize decompression stream: %s\n", ZSTD_getErrorN= ame(ret)); - return -1; - } - - if (!level) - return 0; - - data->cstream =3D ZSTD_createCStream(); - if (data->cstream =3D=3D NULL) { - pr_err("Couldn't create compression stream.\n"); - return -1; - } - - ret =3D ZSTD_initCStream(data->cstream, level); - if (ZSTD_isError(ret)) { - pr_err("Failed to initialize compression stream: %s\n", ZSTD_getErrorNam= e(ret)); - return -1; - } - + data->comp_level =3D level; + data->dstream =3D NULL; + data->cstream =3D NULL; return 0; } =20 @@ -54,7 +28,7 @@ int zstd_fini(struct zstd_data *data) return 0; } =20 -size_t zstd_compress_stream_to_records(struct zstd_data *data, void *dst, = size_t dst_size, +ssize_t zstd_compress_stream_to_records(struct zstd_data *data, void *dst,= size_t dst_size, void *src, size_t src_size, size_t max_record_size, size_t process_header(void *record, size_t increment)) { @@ -63,6 +37,21 @@ size_t zstd_compress_stream_to_records(struct zstd_data = *data, void *dst, size_t ZSTD_outBuffer output; void *record; =20 + if (!data->cstream) { + data->cstream =3D ZSTD_createCStream(); + if (data->cstream =3D=3D NULL) { + pr_err("Couldn't create compression stream.\n"); + return -1; + } + + ret =3D ZSTD_initCStream(data->cstream, data->comp_level); + if (ZSTD_isError(ret)) { + pr_err("Failed to initialize compression stream: %s\n", + ZSTD_getErrorName(ret)); + return -1; + } + } + while (input.pos < input.size) { record =3D dst; size =3D process_header(record, 0); @@ -96,6 +85,20 @@ size_t zstd_decompress_stream(struct zstd_data *data, vo= id *src, size_t src_size ZSTD_inBuffer input =3D { src, src_size, 0 }; ZSTD_outBuffer output =3D { dst, dst_size, 0 }; =20 + if (!data->dstream) { + data->dstream =3D ZSTD_createDStream(); + if (data->dstream =3D=3D NULL) { + pr_err("Couldn't create decompression stream.\n"); + return 0; + } + + ret =3D ZSTD_initDStream(data->dstream); + if (ZSTD_isError(ret)) { + pr_err("Failed to initialize decompression stream: %s\n", + ZSTD_getErrorName(ret)); + return 0; + } + } while (input.pos < input.size) { ret =3D ZSTD_decompressStream(data->dstream, &output, &input); if (ZSTD_isError(ret)) { --=20 2.42.0.758.gaed0368e0e-goog