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=-0.8 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SPF_PASS,URIBL_BLOCKED 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 1467EC43142 for ; Tue, 26 Jun 2018 17:25:20 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id C25CC2642C for ; Tue, 26 Jun 2018 17:25:19 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org C25CC2642C Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933563AbeFZRZS (ORCPT ); Tue, 26 Jun 2018 13:25:18 -0400 Received: from mga05.intel.com ([192.55.52.43]:9093 "EHLO mga05.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932884AbeFZRZQ (ORCPT ); Tue, 26 Jun 2018 13:25:16 -0400 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga006.jf.intel.com ([10.7.209.51]) by fmsmga105.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 26 Jun 2018 10:25:16 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.51,275,1526367600"; d="scan'208";a="53071846" Received: from linux.intel.com ([10.54.29.200]) by orsmga006.jf.intel.com with ESMTP; 26 Jun 2018 10:25:16 -0700 Received: from [10.252.9.225] (abudanko-mobl.ccr.corp.intel.com [10.252.9.225]) by linux.intel.com (Postfix) with ESMTP id AE5C758011E; Tue, 26 Jun 2018 10:25:11 -0700 (PDT) Subject: Re: [RFC 3/4] perf: Allow per PMU access control To: Tvrtko Ursulin , linux-kernel@vger.kernel.org Cc: Tvrtko Ursulin , Thomas Gleixner , Peter Zijlstra , Ingo Molnar , "H. Peter Anvin" , Arnaldo Carvalho de Melo , Alexander Shishkin , Jiri Olsa , Namhyung Kim , Madhavan Srinivasan , Andi Kleen , x86@kernel.org, Alexey Budankov References: <20180626153642.5587-1-tvrtko.ursulin@linux.intel.com> <20180626153642.5587-4-tvrtko.ursulin@linux.intel.com> From: Alexey Budankov Message-ID: Date: Tue, 26 Jun 2018 20:25:10 +0300 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.8.0 MIME-Version: 1.0 In-Reply-To: <20180626153642.5587-4-tvrtko.ursulin@linux.intel.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi, On 26.06.2018 18:36, Tvrtko Ursulin wrote: > From: Tvrtko Ursulin > > For situations where sysadmins might want to allow different level of > access control for different PMUs, we start creating per-PMU > perf_event_paranoid controls in sysfs. > > These work in equivalent fashion as the existing perf_event_paranoid > sysctl, which now becomes the parent control for each PMU. > > On PMU registration the global/parent value will be inherited by each PMU, > as it will be propagated to all registered PMUs when the sysctl is > updated. > > At any later point individual PMU access controls, located in > /device//perf_event_paranoid, can be adjusted to achieve > fine grained access control. > > Signed-off-by: Tvrtko Ursulin > Cc: Thomas Gleixner > Cc: Peter Zijlstra > Cc: Ingo Molnar > Cc: "H. Peter Anvin" > Cc: Arnaldo Carvalho de Melo > Cc: Alexander Shishkin > Cc: Jiri Olsa > Cc: Namhyung Kim > Cc: Madhavan Srinivasan > Cc: Andi Kleen > Cc: Alexey Budankov > Cc: linux-kernel@vger.kernel.org > Cc: x86@kernel.org > --- > include/linux/perf_event.h | 12 ++++++-- > kernel/events/core.c | 59 ++++++++++++++++++++++++++++++++++++++ > kernel/sysctl.c | 4 ++- > 3 files changed, 71 insertions(+), 4 deletions(-) > > diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h > index d7938d88c028..22e91cc2d9e1 100644 > --- a/include/linux/perf_event.h > +++ b/include/linux/perf_event.h > @@ -271,6 +271,9 @@ struct pmu { > /* number of address filters this PMU can do */ > unsigned int nr_addr_filters; > > + /* per PMU access control */ > + int perf_event_paranoid; It looks like it needs to be declared as atomic and atomic_read/atomic_write operations need to be explicitly used below in the patch as far this variable may be manipulated by different threads at the same time without explicit locking. > + > /* > * Fully disable/enable this PMU, can be used to protect from the PMI > * as well as for lazy/batch writing of the MSRs. > @@ -1168,6 +1171,9 @@ extern int sysctl_perf_cpu_time_max_percent; > > extern void perf_sample_event_took(u64 sample_len_ns); > > +extern int perf_proc_paranoid_handler(struct ctl_table *table, int write, > + void __user *buffer, size_t *lenp, > + loff_t *ppos); > extern int perf_proc_update_handler(struct ctl_table *table, int write, > void __user *buffer, size_t *lenp, > loff_t *ppos); > @@ -1180,17 +1186,17 @@ int perf_event_max_stack_handler(struct ctl_table *table, int write, > > static inline bool perf_paranoid_tracepoint_raw(const struct pmu *pmu) > { > - return sysctl_perf_event_paranoid > -1; > + return pmu->perf_event_paranoid > -1; > } > > static inline bool perf_paranoid_cpu(const struct pmu *pmu) > { > - return sysctl_perf_event_paranoid > 0; > + return pmu->perf_event_paranoid > 0; > } > > static inline bool perf_paranoid_kernel(const struct pmu *pmu) > { > - return sysctl_perf_event_paranoid > 1; > + return pmu->perf_event_paranoid > 1; > } > > extern void perf_event_init(void); > diff --git a/kernel/events/core.c b/kernel/events/core.c > index 370c89e81722..da36317dc8dc 100644 > --- a/kernel/events/core.c > +++ b/kernel/events/core.c > @@ -432,6 +432,24 @@ static void update_perf_cpu_limits(void) > > static bool perf_rotate_context(struct perf_cpu_context *cpuctx); > > +int perf_proc_paranoid_handler(struct ctl_table *table, int write, > + void __user *buffer, size_t *lenp, > + loff_t *ppos) > +{ > + int ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); > + struct pmu *pmu; > + > + if (ret || !write) > + return ret; > + > + mutex_lock(&pmus_lock); > + list_for_each_entry(pmu, &pmus, entry) > + pmu->perf_event_paranoid = sysctl_perf_event_paranoid; > + mutex_unlock(&pmus_lock); > + > + return 0; > +} > + > int perf_proc_update_handler(struct ctl_table *table, int write, > void __user *buffer, size_t *lenp, > loff_t *ppos) > @@ -9425,6 +9443,41 @@ static void free_pmu_context(struct pmu *pmu) > mutex_unlock(&pmus_lock); > } > > +/* > + * Fine-grained access control: > + */ > +static ssize_t > +perf_event_paranoid_show(struct device *dev, > + struct device_attribute *attr, > + char *page) > +{ > + struct pmu *pmu = dev_get_drvdata(dev); > + > + return snprintf(page, PAGE_SIZE - 1, "%d\n", pmu->perf_event_paranoid); > +} > + > +static ssize_t > +perf_event_paranoid_store(struct device *dev, > + struct device_attribute *attr, > + const char *buf, size_t count) > +{ > + struct pmu *pmu = dev_get_drvdata(dev); > + int ret, val; > + > + ret = kstrtoint(buf, 0, &val); > + if (ret) > + return ret; > + > + if (val < -1 || val > 2) > + return -EINVAL; > + > + pmu->perf_event_paranoid = val; > + > + return count; > +} > + > +static DEVICE_ATTR_RW(perf_event_paranoid); > + > /* > * Let userspace know that this PMU supports address range filtering: > */ > @@ -9539,6 +9592,11 @@ static int pmu_dev_alloc(struct pmu *pmu) > if (ret) > goto free_dev; > > + /* Add fine-grained access control attribute. */ > + ret = device_create_file(pmu->dev, &dev_attr_perf_event_paranoid); > + if (ret) > + goto del_dev; > + > /* For PMUs with address filters, throw in an extra attribute: */ > if (pmu->nr_addr_filters) > ret = device_create_file(pmu->dev, &dev_attr_nr_addr_filters); > @@ -9570,6 +9628,7 @@ int perf_pmu_register(struct pmu *pmu, const char *name, int type) > if (!pmu->pmu_disable_count) > goto unlock; > > + pmu->perf_event_paranoid = sysctl_perf_event_paranoid; > pmu->type = -1; > if (!name) > goto skip_type; > diff --git a/kernel/sysctl.c b/kernel/sysctl.c > index 2d9837c0aff4..7f6fccb64a30 100644 > --- a/kernel/sysctl.c > +++ b/kernel/sysctl.c > @@ -1142,7 +1142,9 @@ static struct ctl_table kern_table[] = { > .data = &sysctl_perf_event_paranoid, > .maxlen = sizeof(sysctl_perf_event_paranoid), > .mode = 0644, > - .proc_handler = proc_dointvec, > + .proc_handler = perf_proc_paranoid_handler, > + .extra1 = &neg_one, > + .extra2 = &two, > }, > { > .procname = "perf_event_mlock_kb", >