From patchwork Tue Oct 18 15:16:30 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Valentin Schneider X-Patchwork-Id: 13010754 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 054ACC4332F for ; Tue, 18 Oct 2022 15:17:16 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229832AbiJRPRP (ORCPT ); Tue, 18 Oct 2022 11:17:15 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43148 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229911AbiJRPRN (ORCPT ); Tue, 18 Oct 2022 11:17:13 -0400 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id D270FD0CCA for ; Tue, 18 Oct 2022 08:17:07 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1666106226; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=SUDmAUGydtZlR/kOcTvw5Xum4UC/l34yYUVzM6Xginc=; b=iGwmUSwiguiOpwL0EHkckh5jMzXUuNiY89Xl+p/c5GlJU+M5FP8W9Svu/vz/slGOueQp/z +QEBH1eH4rYo25P1PWKMPMiMJi46MFdgiZSq4sLoURSx9r+qpDFOs7mKguN4NzhGHKDynK JrSO8bg5qCUrf66w5xTH5ylNescaPEw= Received: from mail-qt1-f197.google.com (mail-qt1-f197.google.com [209.85.160.197]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_128_GCM_SHA256) id us-mta-624-gQsSr3esNG-9KUr_3V-mog-1; Tue, 18 Oct 2022 11:17:04 -0400 X-MC-Unique: gQsSr3esNG-9KUr_3V-mog-1 Received: by mail-qt1-f197.google.com with SMTP id d12-20020a05622a100c00b0039ce6373d2cso4905926qte.9 for ; Tue, 18 Oct 2022 08:17:04 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=SUDmAUGydtZlR/kOcTvw5Xum4UC/l34yYUVzM6Xginc=; b=KfZukkKyeIH3Ih5yAsFLMKwRKhOogx3daiLrIzYyE9Co03IuVnO0vdVldgWLeUhOpi Hwom9C4snQqJpCXKcz35FcOJJMOh6uCQBCGzlgRqK7fYbaj7xMinkaSd38pxVpSpcrwd DHA3nBcSyeNhVTdsWXk9F5bCrn25Ro4X9NHKvK5eB/8Z+C30u2hkhEfjZMVPJhOpw23K 3NrJNwyfeiMLVfqktiUCYt/U1BOLU+8DhEDkYxuus6AyOkiajwEKfgoRh6gKpuGCnLwo gApxOUmmgNfRotmX9pDfPGt5viPuduW+zoQM6p8Mlxigj9EkqKcgNIANUvuYcAHBmSXH pyzQ== X-Gm-Message-State: ACrzQf1GK/w1xzPSsRd8pPRrFYKDwHGxGZ5cH60Njtw2lszYqMx8ADpN plZ0/WVHwDVcbQReyydXJeJVqDUiyHMUH3Xey8GEyrpKgc11qJhXkwts+utWwIIS3S3nnY582kC E6Iw9AAZm4dHBrduMnYDz56f1VbZYxxONcKsVbu8NSDzU8lENR9ESo3EzoZAGjUBnhXjF68Pj+t b+0Lc6v0Y= X-Received: by 2002:a05:620a:404c:b0:6da:ecec:a072 with SMTP id i12-20020a05620a404c00b006daececa072mr2255411qko.358.1666106223592; Tue, 18 Oct 2022 08:17:03 -0700 (PDT) X-Google-Smtp-Source: AMsMyM5r+aNUUY8XbQTfRVTKvES+EGzp9Y99KEtOCO9QME3/qNyaWJWhknQH59I5ICn/7xoIeE0xoQ== X-Received: by 2002:a05:620a:404c:b0:6da:ecec:a072 with SMTP id i12-20020a05620a404c00b006daececa072mr2255379qko.358.1666106223196; Tue, 18 Oct 2022 08:17:03 -0700 (PDT) Received: from vschneid.remote.csb ([149.71.65.94]) by smtp.gmail.com with ESMTPSA id s23-20020a05622a1a9700b003988b3d5280sm2159646qtc.70.2022.10.18.08.17.01 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 18 Oct 2022 08:17:02 -0700 (PDT) From: Valentin Schneider To: linux-trace-devel@vger.kernel.org Cc: Steven Rostedt , Daniel Bristot de Oliveira , Clark Williams , Douglas RAILLARD Subject: [PATCH 2/2] libtraceevent: Pretty-print cpumask fields as a cpulist Date: Tue, 18 Oct 2022 16:16:30 +0100 Message-Id: <20221018151630.1513535-3-vschneid@redhat.com> X-Mailer: git-send-email 2.31.1 In-Reply-To: <20221018151630.1513535-1-vschneid@redhat.com> References: <20221018151630.1513535-1-vschneid@redhat.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-trace-devel@vger.kernel.org Now that we can denote which bitmasks are cpumasks, it makes sense to pretty-print them to a more user-friendly format: a cpulist. There's two hurdles to that: 1) Estimating the required string buffer size. I've tried to condense it down to an estimator function that is computationally simple enough, though it overestimates by ~1/3. For reference, this estimates: 180 bytes for NR_CPUS=64 (x86 defconfig) 911 bytes for NR_CPUS=256 (arm64 defconfig) 2) Iterating through the bits and bytes. The kernel has a collection of carefully crafted bitmask iterators which make this relatively simple (cf. bitmap_list_string()), but I didn't feel justified in importing half a dozen helpers just for one function. I've implemented a "homegrown" byte-parsing logic which isn't the fastest, but is at least condensed to a single function. Signed-off-by: Valentin Schneider --- src/event-parse.c | 168 +++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 167 insertions(+), 1 deletion(-) diff --git a/src/event-parse.c b/src/event-parse.c index 362720c..6922d5b 100644 --- a/src/event-parse.c +++ b/src/event-parse.c @@ -4454,6 +4454,161 @@ static void print_bitmask_to_seq(struct tep_handle *tep, free(str); } +#define log10(n) \ +( \ + n < 10UL ? 0 : \ + n < 100UL ? 1 : \ + n < 1000UL ? 2 : \ + n < 10000UL ? 3 : \ + n < 100000UL ? 4 : \ + n < 1000000UL ? 5 : \ + n < 10000000UL ? 6 : \ + n < 100000000UL ? 7 : \ + n < 1000000000UL ? 8 : \ + 9 \ +) + +/* ilog10(0) should be 1 but the 0 simplifies below math */ +#define ilog10(n) \ +( \ + n == 0 ? 0UL : \ + n == 1 ? 10UL : \ + n == 2 ? 100UL : \ + n == 3 ? 1000UL : \ + n == 4 ? 10000UL : \ + n == 5 ? 100000UL : \ + n == 6 ? 1000000UL : \ + n == 7 ? 10000000UL : \ + n == 8 ? 100000000UL : \ + 1000000000UL \ +) + +static unsigned int cpumask_worst_size(unsigned int nr_bits) +{ + /* + * Printing all the CPUs separated by a comma is a decent bound for the + * maximum memory required to print a cpumask (a slightly better bound + * is chunks of 2 bits set, i.e. 0-1,3-4,6-7...). + * + * e.g. for nr_bits=132: + * - 131 commas + * - 10 * 1 chars for CPUS [0, 9] + * - 90 * 2 chars for CPUS [10-99] + * - 32 * 3 chars for CPUS [100-131] + */ + unsigned int last_cpu = nr_bits - 1; + unsigned int nr_chars = nr_bits - 1; + int last_lvl = log10(last_cpu); + + /* All log10 levels before the last one have all values used */ + for (int lvl = 0; lvl < last_lvl; lvl++) { + int nr_values = ilog10(lvl + 1) - ilog10(lvl); + + nr_chars += nr_values * (lvl + 1); + } + /* Last level is incomplete */ + nr_chars += (nr_bits - ilog10(last_lvl)) * (last_lvl + 1); + + return nr_chars; +} + +static void print_cpumask_to_seq(struct tep_handle *tep, + struct trace_seq *s, const char *format, + int len_arg, const void *data, int size) +{ + int firstone = -1, firstzero = -1; + int nr_bits = size * 8; + bool first = true; + int str_size = 0; + char buf[12]; /* '-' + log10(2^32) + 1 digits + '\0' */ + char *str; + int index; + int i; + + str = malloc(cpumask_worst_size(nr_bits) + 1); + if (!str) { + do_warning("%s: not enough memory!", __func__); + return; + } + + for (i = 0; i < size; i++) { + unsigned char byte; + int fmtsize; + + if (tep->file_bigendian) + index = size - (i + 1); + else + index = i; + + /* Byte by byte scan, not the best... */ + byte = *(((unsigned char *)data) + index); +more: + /* First find a bit set to one...*/ + if (firstone < 0 && byte) { + /* + * Set all lower bits, so a later ffz on this same byte + * is guaranteed to find a later bit. + */ + firstone = ffs(byte) - 1; + byte |= (1 << firstone) - 1; + firstone += i * 8; + } + + if (firstone < 0) + continue; + + /* ...Then find a bit set to zero */ + if ((~byte) & 0xFF) { + /* + * Clear all lower bits, so a later ffs on this same + * byte is guaranteed to find a later bit. + */ + firstzero = ffs(~byte) - 1; + byte &= ~((1 << (firstzero)) - 1); + firstzero += i * 8; + } else if (i == size - 1) { /* ...Or reach the end of the mask */ + firstzero = nr_bits; + byte = 0; + } else { + continue; + } + + /* We've found a bit set to one, and a later bit set to zero. */ + if (!first) { + str[str_size] = ','; + str_size++; + } + first = false; + + /* It takes {log10(number) + 1} chars to format a number */ + fmtsize = log10(firstone) + 1; + snprintf(buf, fmtsize + 1, "%d", firstone); + memcpy(str + str_size, buf, fmtsize); + str_size += fmtsize; + + if (firstzero > firstone + 1) { + fmtsize = log10(firstzero - 1) + 2; + snprintf(buf, fmtsize + 1, "-%d", firstzero - 1); + memcpy(str + str_size, buf, fmtsize); + str_size += fmtsize; + } + + firstzero = firstone = -1; + if (byte) + goto more; + } + + str[str_size] = 0; + str_size++; + + if (len_arg >= 0) + trace_seq_printf(s, format, len_arg, str); + else + trace_seq_printf(s, format, str); + + free(str); +} + static void print_str_arg(struct trace_seq *s, void *data, int size, struct tep_event *event, const char *format, int len_arg, struct tep_print_arg *arg) @@ -4657,7 +4812,6 @@ static void print_str_arg(struct trace_seq *s, void *data, int size, case TEP_PRINT_BSTRING: print_str_to_seq(s, format, len_arg, arg->string.string); break; - case TEP_PRINT_CPUMASK: case TEP_PRINT_BITMASK: { if (!arg->bitmask.field) { arg->bitmask.field = tep_find_any_field(event, arg->bitmask.bitmask); @@ -4670,6 +4824,18 @@ static void print_str_arg(struct trace_seq *s, void *data, int size, data + offset, len); break; } + case TEP_PRINT_CPUMASK: { + if (!arg->bitmask.field) { + arg->bitmask.field = tep_find_any_field(event, arg->bitmask.bitmask); + arg->bitmask.offset = arg->bitmask.field->offset; + } + if (!arg->bitmask.field) + break; + dynamic_offset_field(tep, arg->bitmask.field, data, size, &offset, &len); + print_cpumask_to_seq(tep, s, format, len_arg, + data + offset, len); + break; + } case TEP_PRINT_OP: /* * The only op for string should be ? :