From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755016Ab3BDTD6 (ORCPT ); Mon, 4 Feb 2013 14:03:58 -0500 Received: from mx1.redhat.com ([209.132.183.28]:48142 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754833Ab3BDTD5 (ORCPT ); Mon, 4 Feb 2013 14:03:57 -0500 Date: Mon, 4 Feb 2013 20:02:46 +0100 From: Oleg Nesterov To: Arnaldo Carvalho de Melo , Ingo Molnar , Peter Zijlstra , Srikar Dronamraju , Steven Rostedt Cc: Anton Arapov , Frank Eigler , Jiri Olsa , Josh Stone , Masami Hiramatsu , "Suzuki K. Poulose" , linux-kernel@vger.kernel.org Subject: [PATCH 2/7] perf: Introduce hw_perf_event->tp_target and ->tp_list Message-ID: <20130204190246.GA10865@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20130204190225.GA10840@redhat.com> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org sys_perf_event_open()->perf_init_event(event) is called before find_get_context(event), this means that event->ctx == NULL when class->reg(TRACE_REG_PERF_REGISTER/OPEN) is called and thus it can't know if this event is per-task or system-wide. This patch adds hw_perf_event->tp_target for PERF_TYPE_TRACEPOINT, this is analogous to PERF_TYPE_BREAKPOINT/bp_target we already have. The patch also moves ->bp_target up so that it can overlap with the new member, this can help the compiler to generate the better code. trace_uprobe_register() will use it for prefiltering to avoid the unnecessary breakpoints in mm's we do not want to trace. ->tp_target doesn't have its own reference, but we can rely on the fact that either sys_perf_event_open() holds a reference, or it is equal to event->ctx->task. So this pointer is always valid until free_event(). Also add the "struct list_head tp_list" into this union. It is not strictly necessary, but it can simplify the next changes and we can add it for free. Signed-off-by: Oleg Nesterov --- include/linux/perf_event.h | 9 +++++++-- kernel/events/core.c | 5 ++++- 2 files changed, 11 insertions(+), 3 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 6bfb2fa..c9775e9 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -135,16 +135,21 @@ struct hw_perf_event { struct { /* software */ struct hrtimer hrtimer; }; + struct { /* tracepoint */ + struct task_struct *tp_target; + /* for tp_event->class */ + struct list_head tp_list; + }; #ifdef CONFIG_HAVE_HW_BREAKPOINT struct { /* breakpoint */ - struct arch_hw_breakpoint info; - struct list_head bp_list; /* * Crufty hack to avoid the chicken and egg * problem hw_breakpoint has with context * creation and event initalization. */ struct task_struct *bp_target; + struct arch_hw_breakpoint info; + struct list_head bp_list; }; #endif }; diff --git a/kernel/events/core.c b/kernel/events/core.c index 1b2e516..340fb53 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6162,11 +6162,14 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, if (task) { event->attach_state = PERF_ATTACH_TASK; + + if (attr->type == PERF_TYPE_TRACEPOINT) + event->hw.tp_target = task; #ifdef CONFIG_HAVE_HW_BREAKPOINT /* * hw_breakpoint is a bit difficult here.. */ - if (attr->type == PERF_TYPE_BREAKPOINT) + else if (attr->type == PERF_TYPE_BREAKPOINT) event->hw.bp_target = task; #endif } -- 1.5.5.1