From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.7 required=3.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FORGED_FROMDOMAIN,FREEMAIL_FROM, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id C53BEC4332F for ; Tue, 14 Sep 2021 13:14:51 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id B173D610D1 for ; Tue, 14 Sep 2021 13:14:51 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233218AbhINNQI (ORCPT ); Tue, 14 Sep 2021 09:16:08 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:48872 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S233235AbhINNQH (ORCPT ); Tue, 14 Sep 2021 09:16:07 -0400 Received: from mail-wm1-x332.google.com (mail-wm1-x332.google.com [IPv6:2a00:1450:4864:20::332]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 4950BC061574 for ; Tue, 14 Sep 2021 06:14:50 -0700 (PDT) Received: by mail-wm1-x332.google.com with SMTP id 192-20020a1c04c9000000b002f7a4ab0a49so1978441wme.0 for ; Tue, 14 Sep 2021 06:14:50 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=uYbzqfK8SdLpGcCUm8lVEKofe44vQ9EzJmzia4J/yVs=; b=BX8UjLZpDbobENtr3ULCLQ/zBt1yfD8xCa7xD4v6xwCJSltbRRj5GseHCqfszpvmVL SyHedSadKsydlvXbXZ4ep8rMS3xB0AloA42GB4EJk2sCREWRnfFXNFghUd2RJZQcvxaI t3TvHFUXLOnYBB1o2ckyeaMS9UL1psZnaMp0FnqQ9Q2Gyg6uswhTDn+p+h37NOeVya23 6G//YdR8znzvm/ZdxX+lpE2xVCc5CJvY/F4js9RfY6aTjKhO9ZqwLrP1zzWeb9x8vifs 55CBikiNiMrVAe9MZ6fEqS0+FVts4XL+1GF1aiWPiDC/FQc/d5oz3/IjhGurw2hPMt7y DNeQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=uYbzqfK8SdLpGcCUm8lVEKofe44vQ9EzJmzia4J/yVs=; b=mHm4CDanabfxuDym0uajkzHxuZZ4KnE4oNlYNIxQKhQtjo3hybB79ozsMCUsT01eJr 3CFpFW5gMpj3ZEYCXStoUM+nu2UIh4lk9QEIiArZr/3N8H9cPrHHgv/9ZP0bJLczUOza xjPEmX9yGQUGR60wC8r6wUdYs2AshrG1+Kv32xlr5jkn7w7LRbQEi0mZpY19IZqko/gU oy9Usg9vJXaBvOPtd66qozfIUjhTQQ7lwQ7xEnUvy88XIXh27ErQYjJQU86T/gvRJVvc KG1sZGC6bM/qkoNyj9qB+urcYMwN5yywBrgewvCskl/4j4SGSPV92Trm6OfQdmk0Vufa 5qWA== X-Gm-Message-State: AOAM5320UhR/GWtT4MAHpmncSeCNQGcIw4jP5tOV/Y2uPvvbyr6HI9cE kM/SE6WRB5Eo3onhOIDwwhDgdTkqz3c= X-Google-Smtp-Source: ABdhPJxNnP20g/ySdq8CUx8VKSD7unM3p+3uaPTMsyajMQ7XgS/+9bjNy4tUuUbxGn47r6Qs7vaorQ== X-Received: by 2002:a1c:1fd3:: with SMTP id f202mr2161490wmf.44.1631625288905; Tue, 14 Sep 2021 06:14:48 -0700 (PDT) Received: from oberon.zico.biz ([83.222.187.186]) by smtp.gmail.com with ESMTPSA id y9sm1274760wmj.36.2021.09.14.06.14.47 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 14 Sep 2021 06:14:48 -0700 (PDT) From: "Tzvetomir Stoyanov (VMware)" To: rostedt@goodmis.org Cc: linux-trace-devel@vger.kernel.org Subject: [PATCH v3 08/25] trace-cmd library: Move CPU flyrecord trace metadata into the buffer option, for trace file version 7 Date: Tue, 14 Sep 2021 16:14:21 +0300 Message-Id: <20210914131438.3965484-9-tz.stoyanov@gmail.com> X-Mailer: git-send-email 2.31.1 In-Reply-To: <20210914131438.3965484-1-tz.stoyanov@gmail.com> References: <20210914131438.3965484-1-tz.stoyanov@gmail.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: linux-trace-devel@vger.kernel.org Extended the BUFFER trace option in trace file version 7 with CPU flyrecord trace metadata. In the version 6 of the trace file, this metadata is located at several places in the file. Part of the metadata is only for the top trace instance, thus limiting per-instance configuration. Moving all CPU trace related metadata in the BUFFER option simplifies the parsing and makes per-instance configuration more flexible. In the new file structure, the top instance is treated as any other instances. The format of the extended BUFFER option is: - offset of the buffer in the trace file - name of the buffer - trace clock, used in this buffer for events timestamps - count of CPUs with trace data - array, describing each CPU with trace data: - CPU id - offset of CPU trace data in the trace file - size of the recorded CPU trace data Signed-off-by: Tzvetomir Stoyanov (VMware) --- lib/trace-cmd/trace-output.c | 187 +++++++++++++++++++++++++---------- 1 file changed, 137 insertions(+), 50 deletions(-) diff --git a/lib/trace-cmd/trace-output.c b/lib/trace-cmd/trace-output.c index 8bc995c7..e2d03f1c 100644 --- a/lib/trace-cmd/trace-output.c +++ b/lib/trace-cmd/trace-output.c @@ -1589,7 +1589,7 @@ int tracecmd_append_options(struct tracecmd_output *handle) } static struct tracecmd_option * -add_buffer_option(struct tracecmd_output *handle, const char *name, int cpus) +add_buffer_option_v6(struct tracecmd_output *handle, const char *name, int cpus) { struct tracecmd_option *option; char *buf; @@ -1639,8 +1639,11 @@ int tracecmd_write_buffer_info(struct tracecmd_output *handle) struct tracecmd_option *option; struct tracecmd_buffer *buf; + if (HAS_SECTIONS(handle)) + return 0; + list_for_each_entry(buf, &handle->buffers, list) { - option = add_buffer_option(handle, buf->name, buf->cpus); + option = add_buffer_option_v6(handle, buf->name, buf->cpus); if (!option) return -1; buf->option = option; @@ -1691,6 +1694,98 @@ int tracecmd_write_cmdlines(struct tracecmd_output *handle) return 0; } +static char *get_clock(struct tracecmd_output *handle) +{ + struct tracefs_instance *inst; + + if (handle->trace_clock) + return handle->trace_clock; + + /* + * If no clock is set on this handle, get the trace clock of + * the top instance in the handle's tracing dir + */ + inst = tracefs_instance_alloc(handle->tracing_dir, NULL); + if (!inst) + return NULL; + handle->trace_clock = tracefs_get_clock(inst); + tracefs_instance_free(inst); + return handle->trace_clock; +} + +__hidden struct tracecmd_option * +out_add_buffer_option_v7(struct tracecmd_output *handle, const char *name, + unsigned short id, unsigned long long data_offset, + int cpus, struct data_file_write *cpu_data) +{ + struct tracecmd_option *option; + int i, j = 0, k = 0; + int *cpu_ids = NULL; + struct iovec *vect; + char *clock; + + if (!HAS_SECTIONS(handle)) + return NULL; + + clock = get_clock(handle); + + /* Buffer flyrecord option, v7: + * - trace data offset in the file + * - buffer name + * - buffer clock + * - CPU count + * - for each CPU: + * - CPU id + * - CPU trace data offset in the file + * - CPU trace data size + */ + + /* Buffer latency option, v7: + * - trace data offset in the file + * - buffer name + * - buffer clock + */ + + vect = calloc(5 + (cpus * 3), sizeof(struct iovec)); + if (!vect) + return NULL; + if (cpus) { + cpu_ids = calloc(cpus, sizeof(int)); + if (!cpu_ids) { + free(vect); + return NULL; + } + } + vect[j].iov_base = (void *) &data_offset; + vect[j++].iov_len = 8; + vect[j].iov_base = (void *) name; + vect[j++].iov_len = strlen(name) + 1; + vect[j].iov_base = (void *) clock; + vect[j++].iov_len = strlen(clock) + 1; + if (id == TRACECMD_OPTION_BUFFER) { + vect[j].iov_base = (void *) &k; + vect[j++].iov_len = 4; + for (i = 0; i < cpus; i++) { + if (!cpu_data[i].file_size) + continue; + cpu_ids[i] = i; + vect[j].iov_base = &cpu_ids[i]; + vect[j++].iov_len = 4; + vect[j].iov_base = &cpu_data[i].data_offset; + vect[j++].iov_len = 8; + vect[j].iov_base = &cpu_data[i].write_size; + vect[j++].iov_len = 8; + k++; + } + } + + option = tracecmd_add_option_v(handle, id, vect, j); + free(vect); + free(cpu_ids); + + return option; +} + struct tracecmd_output *tracecmd_create_file_latency(const char *output_file, int cpus) { struct tracecmd_output *handle; @@ -1767,8 +1862,8 @@ out: return ret; } -static int update_buffer_cpu_offset(struct tracecmd_output *handle, - const char *name, tsize_t offset) +static int update_buffer_cpu_offset_v6(struct tracecmd_output *handle, + const char *name, tsize_t offset) { tsize_t b_offset; tsize_t current; @@ -1797,26 +1892,6 @@ static int update_buffer_cpu_offset(struct tracecmd_output *handle, return 0; } - -static char *get_clock(struct tracecmd_output *handle) -{ - struct tracefs_instance *inst; - - if (handle->trace_clock) - return handle->trace_clock; - - /* - * If no clock is set on this handle, get the trace clock of - * the top instance in the handle's tracing dir - */ - inst = tracefs_instance_alloc(handle->tracing_dir, NULL); - if (!inst) - return NULL; - handle->trace_clock = tracefs_get_clock(inst); - tracefs_instance_free(inst); - return handle->trace_clock; -} - __hidden int out_write_cpu_data(struct tracecmd_output *handle, int cpus, struct cpu_data_source *data, const char *buff_name) { @@ -1840,7 +1915,7 @@ __hidden int out_write_cpu_data(struct tracecmd_output *handle, } data_offs = do_lseek(handle, 0, SEEK_CUR); - if (do_write_check(handle, "flyrecord", 10)) + if (!HAS_SECTIONS(handle) && do_write_check(handle, "flyrecord", 10)) goto out_free; data_files = calloc(cpus, sizeof(struct data_file_write)); @@ -1850,19 +1925,24 @@ __hidden int out_write_cpu_data(struct tracecmd_output *handle, for (i = 0; i < cpus; i++) { data_files[i].file_size = data[i].size; /* Write 0 for trace data offset and size and store offsets of these fields */ - endian8 = 0; - data_files[i].doffset = do_lseek(handle, 0, SEEK_CUR); - if (do_write_check(handle, &endian8, 8)) - goto out_free; - data_files[i].soffset = do_lseek(handle, 0, SEEK_CUR); - if (do_write_check(handle, &endian8, 8)) + if (!HAS_SECTIONS(handle)) { + endian8 = 0; + data_files[i].doffset = do_lseek(handle, 0, SEEK_CUR); + if (do_write_check(handle, &endian8, 8)) + goto out_free; + data_files[i].soffset = do_lseek(handle, 0, SEEK_CUR); + if (do_write_check(handle, &endian8, 8)) + goto out_free; + } + } + + if (!HAS_SECTIONS(handle)) { + update_buffer_cpu_offset_v6(handle, buff_name, data_offs); + clock = get_clock(handle); + if (clock && save_clock(handle, clock)) goto out_free; } - update_buffer_cpu_offset(handle, buff_name, data_offs); - clock = get_clock(handle); - if (clock && save_clock(handle, clock)) - goto out_free; for (i = 0; i < cpus; i++) { data_files[i].data_offset = do_lseek(handle, 0, SEEK_CUR); /* Page align offset */ @@ -1888,26 +1968,33 @@ __hidden int out_write_cpu_data(struct tracecmd_output *handle, data_files[i].write_size = 0; } - /* Write the real CPU data offset in the file */ - if (do_lseek(handle, data_files[i].doffset, SEEK_SET) == (off64_t)-1) - goto out_free; - endian8 = convert_endian_8(handle, data_files[i].data_offset); - if (do_write_check(handle, &endian8, 8)) - goto out_free; - /* Write the real CPU data size in the file */ - if (do_lseek(handle, data_files[i].soffset, SEEK_SET) == (off64_t)-1) - goto out_free; - endian8 = convert_endian_8(handle, data_files[i].write_size); - if (do_write_check(handle, &endian8, 8)) - goto out_free; - offset = data_files[i].data_offset + data_files[i].write_size; - if (do_lseek(handle, offset, SEEK_SET) == (off64_t)-1) - goto out_free; + if (!HAS_SECTIONS(handle)) { + /* Write the real CPU data offset in the file */ + if (do_lseek(handle, data_files[i].doffset, SEEK_SET) == (off64_t)-1) + goto out_free; + endian8 = convert_endian_8(handle, data_files[i].data_offset); + if (do_write_check(handle, &endian8, 8)) + goto out_free; + /* Write the real CPU data size in the file */ + if (do_lseek(handle, data_files[i].soffset, SEEK_SET) == (off64_t)-1) + goto out_free; + endian8 = convert_endian_8(handle, data_files[i].write_size); + if (do_write_check(handle, &endian8, 8)) + goto out_free; + offset = data_files[i].data_offset + data_files[i].write_size; + if (do_lseek(handle, offset, SEEK_SET) == (off64_t)-1) + goto out_free; + } if (!tracecmd_get_quiet(handle)) fprintf(stderr, " %llu bytes in size\n", (unsigned long long)data_files[i].write_size); } + if (HAS_SECTIONS(handle) && + !out_add_buffer_option_v7(handle, buff_name, + TRACECMD_OPTION_BUFFER, data_offs, cpus, data_files)) + goto out_free; + if (do_lseek(handle, 0, SEEK_END) == (off64_t)-1) goto out_free; -- 2.31.1