From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from desiato.infradead.org (desiato.infradead.org [90.155.92.199]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 60AC818AE1 for ; Mon, 12 Jun 2023 09:39:50 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=infradead.org; s=desiato.20200630; h=Content-Type:MIME-Version:References: Subject:Cc:To:From:Date:Message-ID:Sender:Reply-To:Content-Transfer-Encoding: Content-ID:Content-Description:In-Reply-To; bh=A6Nf2exxWEURXKIPC4PisBIYNsQ8WOBsy7I8/r8MB2I=; b=QBEQojlPOctoq6m4at+zDWspSt wLwIrtYUbLj0j1sfpEmxBViAyHgcgROxzxQwfT3lTGgty+s5tbZ6EbtiB7F+9gLvV8EMLMSghl8I0 AyNN/6oOlyqNToFGZ/ozYQ8tFNBsZrc+DoAmXaDvDSzbp6ZX9mFEI8zG2Xwti2PhIuhoI5h86pVCY KP/0l64QOPPmXvxDqK3cLtzDrPThkS7WcH8QNrScYz1W2aZRD1jsA+X01l5cHGCM6Baf9JIwMZxHw SZiz3drbNw5O0sfshDo9uroFr2tGF7VKAZaQxZnbY9B8qhuf6oZuiJ+ynC2QC/5htAoiB4EkKK1k4 8gpd0qqA==; Received: from j130084.upc-j.chello.nl ([24.132.130.84] helo=noisy.programming.kicks-ass.net) by desiato.infradead.org with esmtpsa (Exim 4.96 #2 (Red Hat Linux)) id 1q8e0l-008kQV-1I; Mon, 12 Jun 2023 09:38:56 +0000 Received: from hirez.programming.kicks-ass.net (hirez.programming.kicks-ass.net [192.168.1.225]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits)) (Client did not present a certificate) by noisy.programming.kicks-ass.net (Postfix) with ESMTPS id 0E6EC302E28; Mon, 12 Jun 2023 11:38:53 +0200 (CEST) Received: by hirez.programming.kicks-ass.net (Postfix, from userid 0) id 725F830A77B5F; Mon, 12 Jun 2023 11:38:48 +0200 (CEST) Message-ID: <20230612093539.085862001@infradead.org> User-Agent: quilt/0.66 Date: Mon, 12 Jun 2023 11:07:35 +0200 From: Peter Zijlstra To: torvalds@linux-foundation.org, keescook@chromium.org, gregkh@linuxfoundation.org, pbonzini@redhat.com Cc: masahiroy@kernel.org, nathan@kernel.org, ndesaulniers@google.com, nicolas@fjasle.eu, catalin.marinas@arm.com, will@kernel.org, vkoul@kernel.org, trix@redhat.com, ojeda@kernel.org, peterz@infradead.org, mingo@redhat.com, longman@redhat.com, boqun.feng@gmail.com, dennis@kernel.org, tj@kernel.org, cl@linux.com, acme@kernel.org, mark.rutland@arm.com, alexander.shishkin@linux.intel.com, jolsa@kernel.org, namhyung@kernel.org, irogers@google.com, adrian.hunter@intel.com, juri.lelli@redhat.com, vincent.guittot@linaro.org, dietmar.eggemann@arm.com, rostedt@goodmis.org, bsegall@google.com, mgorman@suse.de, bristot@redhat.com, vschneid@redhat.com, paulmck@kernel.org, frederic@kernel.org, quic_neeraju@quicinc.com, joel@joelfernandes.org, josh@joshtriplett.org, mathieu.desnoyers@efficios.com, jiangshanlai@gmail.com, rientjes@google.com, vbabka@suse.cz, roman.gushchin@linux.dev, 42.hyeyoo@gmail.com, apw@canonical.com, joe@perches.com, dwaipayanray1@gmail.com, lukas.bulwahn@gmail.com, john.johansen@canonical.com, paul@paul-moore.com, jmorris@namei.org, serge@hallyn.com, linux-kbuild@vger.kernel.org, linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, llvm@lists.linux.dev, linux-perf-users@vger.kernel.org, rcu@vger.kernel.org, linux-security-module@vger.kernel.org, tglx@linutronix.de, ravi.bangoria@amd.com, error27@gmail.com, luc.vanoostenryck@gmail.com Subject: [PATCH v3 22/57] perf: Fix cpuctx refcounting References: <20230612090713.652690195@infradead.org> Precedence: bulk X-Mailing-List: llvm@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Fixes: bd2756811766 ("perf: Rewrite core context handling") Signed-off-by: Peter Zijlstra (Intel) --- include/linux/perf_event.h | 13 ++++++++----- kernel/events/core.c | 16 ++++++++++++++++ 2 files changed, 24 insertions(+), 5 deletions(-) --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -841,11 +841,11 @@ struct perf_event { }; /* - * ,-----------------------[1:n]----------------------. - * V V - * perf_event_context <-[1:n]-> perf_event_pmu_context <--- perf_event - * ^ ^ | | - * `--------[1:n]---------' `-[n:1]-> pmu <-[1:n]-' + * ,-----------------------[1:n]------------------------. + * V V + * perf_event_context <-[1:n]-> perf_event_pmu_context <-[1:n]- perf_event + * | | + * `--[n:1]-> pmu <-[1:n]--' * * * struct perf_event_pmu_context lifetime is refcount based and RCU freed @@ -863,6 +863,9 @@ struct perf_event { * ctx->mutex pinning the configuration. Since we hold a reference on * group_leader (through the filedesc) it can't go away, therefore it's * associated pmu_ctx must exist and cannot change due to ctx->mutex. + * + * perf_event holds a refcount on perf_event_context + * perf_event holds a refcount on perf_event_pmu_context */ struct perf_event_pmu_context { struct pmu *pmu; --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -4809,6 +4809,11 @@ find_get_pmu_context(struct pmu *pmu, st void *task_ctx_data = NULL; if (!ctx->task) { + /* + * perf_pmu_migrate_context() / __perf_pmu_install_event() + * relies on the fact that find_get_pmu_context() cannot fail + * for CPU contexts. + */ struct perf_cpu_pmu_context *cpc; cpc = per_cpu_ptr(pmu->cpu_pmu_context, event->cpu); @@ -12832,6 +12837,13 @@ static void __perf_pmu_install_event(str { struct perf_event_pmu_context *epc; + /* + * Now that the events are unused, put their old ctx and grab a + * reference on the new context. + */ + put_ctx(event->ctx); + get_ctx(ctx); + event->cpu = cpu; epc = find_get_pmu_context(pmu, ctx, event); event->pmu_ctx = epc; @@ -12877,6 +12889,10 @@ void perf_pmu_migrate_context(struct pmu struct perf_event_context *src_ctx, *dst_ctx; LIST_HEAD(events); + /* + * Since per-cpu context is persistent, no need to grab an extra + * reference. + */ src_ctx = &per_cpu_ptr(&perf_cpu_context, src_cpu)->ctx; dst_ctx = &per_cpu_ptr(&perf_cpu_context, dst_cpu)->ctx;