Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751465AbdFZFjR (ORCPT ); Mon, 26 Jun 2017 01:39:17 -0400 Received: from mail-pf0-f171.google.com ([209.85.192.171]:35738 "EHLO mail-pf0-f171.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751414AbdFZFjA (ORCPT ); Mon, 26 Jun 2017 01:39:00 -0400 From: Joel Fernandes To: linux-kernel@vger.kernel.org Cc: mikesart@gmail.com, Joel Fernandes , kernel-team@android.com, Steven Rostedt , Ingo Molnar Subject: [PATCH v4 1/3] tracing: Add support for recording tgid of tasks Date: Sun, 25 Jun 2017 22:38:42 -0700 Message-Id: <20170626053844.5746-2-joelaf@google.com> X-Mailer: git-send-email 2.13.1.611.g7e3b11ae1-goog In-Reply-To: <20170626053844.5746-1-joelaf@google.com> References: <20170626053844.5746-1-joelaf@google.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 15335 Lines: 489 Inorder to support recording of tgid, the following changes are made: * Introduce a new API (tracing_record_taskinfo) to additionally record the tgid along with the task's comm at the same time. This has has the benefit of not setting trace_cmdline_save before all the information for a task is saved. * Add a new API tracing_record_taskinfo_sched_switch to record task information for 2 tasks at a time (previous and next) and use it from sched_switch probe. * Preserve the old API (tracing_record_cmdline) and create it as a wrapper around the new one so that existing callers aren't affected. * Reuse the existing sched_switch and sched_wakeup probes to record tgid information and add a new option 'record-tgid' to enable recording of tgid When record-tgid option isn't enabled to being with, we take care to make sure that there's isn't memory or runtime overhead. Cc: kernel-team@android.com Cc: Steven Rostedt Cc: Ingo Molnar Tested-by: Michael Sartain Signed-off-by: Joel Fernandes --- include/linux/trace_events.h | 13 ++++- kernel/trace/trace.c | 106 ++++++++++++++++++++++++++++++++++---- kernel/trace/trace.h | 9 ++++ kernel/trace/trace_events.c | 42 ++++++++++++++- kernel/trace/trace_sched_switch.c | 75 ++++++++++++++++++++++----- 5 files changed, 219 insertions(+), 26 deletions(-) diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h index a556805eff8a..f73cedfa2e0b 100644 --- a/include/linux/trace_events.h +++ b/include/linux/trace_events.h @@ -151,7 +151,15 @@ trace_event_buffer_lock_reserve(struct ring_buffer **current_buffer, int type, unsigned long len, unsigned long flags, int pc); -void tracing_record_cmdline(struct task_struct *tsk); +#define TRACE_RECORD_CMDLINE BIT(0) +#define TRACE_RECORD_TGID BIT(1) + +void tracing_record_taskinfo(struct task_struct *task, int flags); +void tracing_record_taskinfo_sched_switch(struct task_struct *prev, + struct task_struct *next, int flags); + +void tracing_record_cmdline(struct task_struct *task); +void tracing_record_tgid(struct task_struct *task); int trace_output_call(struct trace_iterator *iter, char *name, char *fmt, ...); @@ -290,6 +298,7 @@ struct trace_subsystem_dir; enum { EVENT_FILE_FL_ENABLED_BIT, EVENT_FILE_FL_RECORDED_CMD_BIT, + EVENT_FILE_FL_RECORDED_TGID_BIT, EVENT_FILE_FL_FILTERED_BIT, EVENT_FILE_FL_NO_SET_FILTER_BIT, EVENT_FILE_FL_SOFT_MODE_BIT, @@ -303,6 +312,7 @@ enum { * Event file flags: * ENABLED - The event is enabled * RECORDED_CMD - The comms should be recorded at sched_switch + * RECORDED_TGID - The tgids should be recorded at sched_switch * FILTERED - The event has a filter attached * NO_SET_FILTER - Set when filter has error and is to be ignored * SOFT_MODE - The event is enabled/disabled by SOFT_DISABLED @@ -315,6 +325,7 @@ enum { enum { EVENT_FILE_FL_ENABLED = (1 << EVENT_FILE_FL_ENABLED_BIT), EVENT_FILE_FL_RECORDED_CMD = (1 << EVENT_FILE_FL_RECORDED_CMD_BIT), + EVENT_FILE_FL_RECORDED_TGID = (1 << EVENT_FILE_FL_RECORDED_TGID_BIT), EVENT_FILE_FL_FILTERED = (1 << EVENT_FILE_FL_FILTERED_BIT), EVENT_FILE_FL_NO_SET_FILTER = (1 << EVENT_FILE_FL_NO_SET_FILTER_BIT), EVENT_FILE_FL_SOFT_MODE = (1 << EVENT_FILE_FL_SOFT_MODE_BIT), diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 1122f151466f..6dec7bef63c7 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -87,7 +87,7 @@ dummy_set_flag(struct trace_array *tr, u32 old_flags, u32 bit, int set) * tracing is active, only save the comm when a trace event * occurred. */ -static DEFINE_PER_CPU(bool, trace_cmdline_save); +static DEFINE_PER_CPU(bool, trace_taskinfo_save); /* * Kill all tracing for good (never come back). @@ -790,7 +790,7 @@ EXPORT_SYMBOL_GPL(tracing_on); static __always_inline void __buffer_unlock_commit(struct ring_buffer *buffer, struct ring_buffer_event *event) { - __this_cpu_write(trace_cmdline_save, true); + __this_cpu_write(trace_taskinfo_save, true); /* If this is the temp buffer, we need to commit fully */ if (this_cpu_read(trace_buffered_event) == event) { @@ -1709,6 +1709,18 @@ void tracing_reset_all_online_cpus(void) } } +int *tgid_map; + +void tracing_alloc_tgid_map(void) +{ + if (tgid_map) + return; + + tgid_map = kzalloc((PID_MAX_DEFAULT + 1) * sizeof(*tgid_map), + GFP_KERNEL); + WARN_ONCE(!tgid_map, "Allocation of tgid_map failed\n"); +} + #define SAVED_CMDLINES_DEFAULT 128 #define NO_CMDLINE_MAP UINT_MAX static arch_spinlock_t trace_cmdline_lock = __ARCH_SPIN_LOCK_UNLOCKED; @@ -1722,7 +1734,7 @@ struct saved_cmdlines_buffer { static struct saved_cmdlines_buffer *savedcmd; /* temporary disable recording */ -static atomic_t trace_record_cmdline_disabled __read_mostly; +static atomic_t trace_record_taskinfo_disabled __read_mostly; static inline char *get_saved_cmdlines(int idx) { @@ -1992,16 +2004,87 @@ void trace_find_cmdline(int pid, char comm[]) preempt_enable(); } -void tracing_record_cmdline(struct task_struct *tsk) +int trace_find_tgid(int pid) +{ + if (unlikely(!tgid_map || !pid || pid > PID_MAX_DEFAULT)) + return 0; + + return tgid_map[pid]; +} + +static int trace_save_tgid(struct task_struct *tsk) +{ + if (unlikely(!tgid_map || !tsk->pid || tsk->pid > PID_MAX_DEFAULT)) + return 0; + + tgid_map[tsk->pid] = tsk->tgid; + return 1; +} + +static bool tracing_record_taskinfo_skip(int flags) +{ + if (unlikely(!(flags & (TRACE_RECORD_CMDLINE | TRACE_RECORD_TGID)))) + return true; + if (atomic_read(&trace_record_taskinfo_disabled) || !tracing_is_on()) + return true; + if (!__this_cpu_read(trace_taskinfo_save)) + return true; + return false; +} + +/** + * tracing_record_taskinfo - record the task info of a task + * + * @task - task to record + * @flags - TRACE_RECORD_CMDLINE for recording comm + * - TRACE_RECORD_TGID for recording tgid + */ +void tracing_record_taskinfo(struct task_struct *task, int flags) +{ + if (tracing_record_taskinfo_skip(flags)) + return; + if ((flags & TRACE_RECORD_CMDLINE) && !trace_save_cmdline(task)) + return; + if ((flags & TRACE_RECORD_TGID) && !trace_save_tgid(task)) + return; + + __this_cpu_write(trace_taskinfo_save, false); +} + +/** + * tracing_record_taskinfo_sched_switch - record task info for sched_switch + * + * @prev - previous task during sched_switch + * @next - next task during sched_switch + * @flags - TRACE_RECORD_CMDLINE for recording comm + * TRACE_RECORD_TGID for recording tgid + */ +void tracing_record_taskinfo_sched_switch(struct task_struct *prev, + struct task_struct *next, int flags) { - if (atomic_read(&trace_record_cmdline_disabled) || !tracing_is_on()) + if (tracing_record_taskinfo_skip(flags)) + return; + + if ((flags & TRACE_RECORD_CMDLINE) && + (!trace_save_cmdline(prev) || !trace_save_cmdline(next))) return; - if (!__this_cpu_read(trace_cmdline_save)) + if ((flags & TRACE_RECORD_TGID) && + (!trace_save_tgid(prev) || !trace_save_tgid(next))) return; - if (trace_save_cmdline(tsk)) - __this_cpu_write(trace_cmdline_save, false); + __this_cpu_write(trace_taskinfo_save, false); +} + +/* Helpers to record a specific task information */ +void tracing_record_cmdline(struct task_struct *task) +{ + tracing_record_taskinfo(task, TRACE_RECORD_CMDLINE); +} + +void tracing_record_tgid(struct task_struct *task) +{ + tracing_record_taskinfo(task, TRACE_RECORD_TGID); } /* @@ -3146,7 +3229,7 @@ static void *s_start(struct seq_file *m, loff_t *pos) #endif if (!iter->snapshot) - atomic_inc(&trace_record_cmdline_disabled); + atomic_inc(&trace_record_taskinfo_disabled); if (*pos != iter->pos) { iter->ent = NULL; @@ -3191,7 +3274,7 @@ static void s_stop(struct seq_file *m, void *p) #endif if (!iter->snapshot) - atomic_dec(&trace_record_cmdline_disabled); + atomic_dec(&trace_record_taskinfo_disabled); trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); @@ -4238,6 +4321,9 @@ int set_tracer_flag(struct trace_array *tr, unsigned int mask, int enabled) if (mask == TRACE_ITER_RECORD_CMD) trace_event_enable_cmd_record(enabled); + if (mask == TRACE_ITER_RECORD_TGID) + trace_event_enable_tgid_record(enabled); + if (mask == TRACE_ITER_EVENT_FORK) trace_event_follow_fork(tr, enabled); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 39fd77330aab..90318b016fd9 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -635,8 +635,13 @@ void trace_graph_return(struct ftrace_graph_ret *trace); int trace_graph_entry(struct ftrace_graph_ent *trace); void set_graph_array(struct trace_array *tr); +extern int *tgid_map; +void tracing_alloc_tgid_map(void); void tracing_start_cmdline_record(void); void tracing_stop_cmdline_record(void); +void tracing_start_tgid_record(void); +void tracing_stop_tgid_record(void); + int register_tracer(struct tracer *type); int is_tracing_stopped(void); @@ -697,6 +702,7 @@ static inline void __trace_stack(struct trace_array *tr, unsigned long flags, extern u64 ftrace_now(int cpu); extern void trace_find_cmdline(int pid, char comm[]); +extern int trace_find_tgid(int pid); extern void trace_event_follow_fork(struct trace_array *tr, bool enable); #ifdef CONFIG_DYNAMIC_FTRACE @@ -1107,6 +1113,7 @@ extern int trace_get_user(struct trace_parser *parser, const char __user *ubuf, C(CONTEXT_INFO, "context-info"), /* Print pid/cpu/time */ \ C(LATENCY_FMT, "latency-format"), \ C(RECORD_CMD, "record-cmd"), \ + C(RECORD_TGID, "record-tgid"), \ C(OVERWRITE, "overwrite"), \ C(STOP_ON_FREE, "disable_on_free"), \ C(IRQ_INFO, "irq-info"), \ @@ -1423,6 +1430,8 @@ struct ftrace_event_field * trace_find_event_field(struct trace_event_call *call, char *name); extern void trace_event_enable_cmd_record(bool enable); +extern void trace_event_enable_tgid_record(bool enable); + extern int event_trace_add_tracer(struct dentry *parent, struct trace_array *tr); extern int event_trace_del_tracer(struct trace_array *tr); diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index e7973e10398c..240c6df95ea6 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -343,6 +343,28 @@ void trace_event_enable_cmd_record(bool enable) mutex_unlock(&event_mutex); } +void trace_event_enable_tgid_record(bool enable) +{ + struct trace_event_file *file; + struct trace_array *tr; + + mutex_lock(&event_mutex); + do_for_each_event_file(tr, file) { + if (!(file->flags & EVENT_FILE_FL_ENABLED)) + continue; + + if (enable) { + tracing_start_tgid_record(); + set_bit(EVENT_FILE_FL_RECORDED_TGID_BIT, &file->flags); + } else { + tracing_stop_tgid_record(); + clear_bit(EVENT_FILE_FL_RECORDED_TGID_BIT, + &file->flags); + } + } while_for_each_event_file(); + mutex_unlock(&event_mutex); +} + static int __ftrace_event_enable_disable(struct trace_event_file *file, int enable, int soft_disable) { @@ -381,6 +403,12 @@ static int __ftrace_event_enable_disable(struct trace_event_file *file, tracing_stop_cmdline_record(); clear_bit(EVENT_FILE_FL_RECORDED_CMD_BIT, &file->flags); } + + if (file->flags & EVENT_FILE_FL_RECORDED_TGID) { + tracing_stop_tgid_record(); + clear_bit(EVENT_FILE_FL_RECORDED_CMD_BIT, &file->flags); + } + call->class->reg(call, TRACE_REG_UNREGISTER, file); } /* If in SOFT_MODE, just set the SOFT_DISABLE_BIT, else clear it */ @@ -407,18 +435,30 @@ static int __ftrace_event_enable_disable(struct trace_event_file *file, } if (!(file->flags & EVENT_FILE_FL_ENABLED)) { + bool cmd = false, tgid = false; /* Keep the event disabled, when going to SOFT_MODE. */ if (soft_disable) set_bit(EVENT_FILE_FL_SOFT_DISABLED_BIT, &file->flags); if (tr->trace_flags & TRACE_ITER_RECORD_CMD) { + cmd = true; tracing_start_cmdline_record(); set_bit(EVENT_FILE_FL_RECORDED_CMD_BIT, &file->flags); } + + if (tr->trace_flags & TRACE_ITER_RECORD_TGID) { + tgid = true; + tracing_start_tgid_record(); + set_bit(EVENT_FILE_FL_RECORDED_TGID_BIT, &file->flags); + } + ret = call->class->reg(call, TRACE_REG_REGISTER, file); if (ret) { - tracing_stop_cmdline_record(); + if (cmd) + tracing_stop_cmdline_record(); + if (tgid) + tracing_stop_tgid_record(); pr_info("event trace: Could not enable event " "%s\n", trace_event_name(call)); break; diff --git a/kernel/trace/trace_sched_switch.c b/kernel/trace/trace_sched_switch.c index 4c896a0101bd..75a0c3accf2a 100644 --- a/kernel/trace/trace_sched_switch.c +++ b/kernel/trace/trace_sched_switch.c @@ -12,27 +12,38 @@ #include "trace.h" -static int sched_ref; +#define RECORD_CMDLINE 1 +#define RECORD_TGID 2 + +static int sched_cmdline_ref; +static int sched_tgid_ref; static DEFINE_MUTEX(sched_register_mutex); static void probe_sched_switch(void *ignore, bool preempt, struct task_struct *prev, struct task_struct *next) { - if (unlikely(!sched_ref)) - return; + int flags; + + flags = (RECORD_TGID * !!sched_tgid_ref) + + (RECORD_CMDLINE * !!sched_cmdline_ref); - tracing_record_cmdline(prev); - tracing_record_cmdline(next); + if (!flags) + return; + tracing_record_taskinfo_sched_switch(prev, next, flags); } static void probe_sched_wakeup(void *ignore, struct task_struct *wakee) { - if (unlikely(!sched_ref)) - return; + int flags; + + flags = (RECORD_TGID * !!sched_tgid_ref) + + (RECORD_CMDLINE * !!sched_cmdline_ref); - tracing_record_cmdline(current); + if (!flags) + return; + tracing_record_taskinfo(current, flags); } static int tracing_sched_register(void) @@ -75,28 +86,64 @@ static void tracing_sched_unregister(void) unregister_trace_sched_wakeup(probe_sched_wakeup, NULL); } -static void tracing_start_sched_switch(void) +static void tracing_start_sched_switch(int ops) { + bool sched_register = (!sched_cmdline_ref && !sched_tgid_ref); mutex_lock(&sched_register_mutex); - if (!(sched_ref++)) + + switch (ops) { + case RECORD_CMDLINE: + sched_cmdline_ref++; + break; + + case RECORD_TGID: + sched_tgid_ref++; + break; + } + + if (sched_tgid_ref > 0 && !tgid_map) + tracing_alloc_tgid_map(); + + if (sched_register && (sched_cmdline_ref || sched_tgid_ref)) tracing_sched_register(); mutex_unlock(&sched_register_mutex); } -static void tracing_stop_sched_switch(void) +static void tracing_stop_sched_switch(int ops) { mutex_lock(&sched_register_mutex); - if (!(--sched_ref)) + + switch (ops) { + case RECORD_CMDLINE: + sched_cmdline_ref--; + break; + + case RECORD_TGID: + sched_tgid_ref--; + break; + } + + if (!sched_cmdline_ref && !sched_tgid_ref) tracing_sched_unregister(); mutex_unlock(&sched_register_mutex); } void tracing_start_cmdline_record(void) { - tracing_start_sched_switch(); + tracing_start_sched_switch(RECORD_CMDLINE); } void tracing_stop_cmdline_record(void) { - tracing_stop_sched_switch(); + tracing_stop_sched_switch(RECORD_CMDLINE); +} + +void tracing_start_tgid_record(void) +{ + tracing_start_sched_switch(RECORD_TGID); +} + +void tracing_stop_tgid_record(void) +{ + tracing_stop_sched_switch(RECORD_TGID); } -- 2.13.1.611.g7e3b11ae1-goog