linux-kernel - Re: [RFC PATCH] tracing: add sched_prio

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 5 Jul 2016 21:50:34 +0000 (UTC)
From:	Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To:	rostedt <rostedt@...dmis.org>
Cc:	Julien Desfossez <jdesfossez@...icios.com>,
	Peter Zijlstra <peterz@...radead.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...nel.org>, daolivei@...hat.com,
	linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH] tracing: add sched_prio_update

----- On Jul 5, 2016, at 11:19 AM, rostedt rostedt@...dmis.org wrote:

> On Mon,  4 Jul 2016 15:46:04 -0400
> Julien Desfossez <jdesfossez@...icios.com> wrote:
> 
> 
>> diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
>> index 9b90c57..fcb0f29 100644
>> --- a/include/trace/events/sched.h
>> +++ b/include/trace/events/sched.h
>> @@ -8,6 +8,34 @@
>>  #include <linux/tracepoint.h>
>>  #include <linux/binfmts.h>
>>  
>> +#define SCHEDULING_POLICY				\
>> +	EM( SCHED_NORMAL,	"SCHED_NORMAL")		\
>> +	EM( SCHED_FIFO,		"SCHED_FIFO")		\
>> +	EM( SCHED_RR,		"SCHED_RR")		\
>> +	EM( SCHED_BATCH,	"SCHED_BATCH")		\
>> +	EM( SCHED_IDLE,		"SCHED_IDLE")		\
>> +	EMe(SCHED_DEADLINE,	"SCHED_DEADLINE")
>> +
>> +/*
>> + * First define the enums in the above macros to be exported to userspace
>> + * via TRACE_DEFINE_ENUM().
>> + */
>> +#undef EM
>> +#undef EMe
>> +#define EM(a, b)	TRACE_DEFINE_ENUM(a);
>> +#define EMe(a, b)	TRACE_DEFINE_ENUM(a);
>> +
>> +SCHEDULING_POLICY
>> +
>> +/*
>> + * Now redefine the EM() and EMe() macros to map the enums to the strings
>> + * that will be printed in the output.
>> + */
>> +#undef EM
>> +#undef EMe
>> +#define EM(a, b)	{a, b},
>> +#define EMe(a, b)	{a, b}
>> +
>>  /*
>>   * Tracepoint for calling kthread_stop, performed to end a kthread:
>>   */
>> @@ -562,6 +590,46 @@ TRACE_EVENT(sched_wake_idle_without_ipi,
>>  
>>  	TP_printk("cpu=%d", __entry->cpu)
>>  );
>> +
>> +/*
>> + * Tracepoint for showing scheduling priority changes.
>> + */
>> +TRACE_EVENT(sched_prio_update,
> 
> I'm fine with the addition of this tracepoint. You'll have to get by
> Peter Zijlstra for it.

Great!

> 
>> +
>> +	TP_PROTO(struct task_struct *tsk),
>> +
>> +	TP_ARGS(tsk),
>> +
>> +	TP_STRUCT__entry(
>> +		__array( char,	comm,	TASK_COMM_LEN	)
> 
> I could imagine this being a high frequency tracepoint, especially with
> a lot of boosting going on. Can we nuke the comm recording and let the
> userspace tools just hook to the sched_switch tracepoint for that?

We can surely do that.

Just to clarify: currently this tracepoint is *not* hooked on PI boosting,
as described in the changelog. This tracepoint is about the prio attributes
set by user-space. The PI boosting temporarily changes the task struct prio
without updating the associated policy, which seems rather
implementation-specific and odd to expose.

Thoughts ?

Thanks,

Mathieu


> 
> -- Steve
> 
> 
>> +		__field( pid_t,	pid			)
>> +		__field( unsigned int,	policy		)
>> +		__field( int,	nice			)
>> +		__field( unsigned int,	rt_priority	)
>> +		__field( u64,	dl_runtime		)
>> +		__field( u64,	dl_deadline		)
>> +		__field( u64,	dl_period		)
>> +	),
>> +
>> +	TP_fast_assign(
>> +		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
>> +		__entry->pid		= tsk->pid;
>> +		__entry->policy		= tsk->policy;
>> +		__entry->nice		= task_nice(tsk);
>> +		__entry->rt_priority	= tsk->rt_priority;
>> +		__entry->dl_runtime	= tsk->dl.dl_runtime;
>> +		__entry->dl_deadline	= tsk->dl.dl_deadline;
>> +		__entry->dl_period	= tsk->dl.dl_period;
>> +	),
>> +
>> +	TP_printk("comm=%s pid=%d, policy=%s, nice=%d, rt_priority=%u, "
>> +			"dl_runtime=%Lu, dl_deadline=%Lu, dl_period=%Lu",
>> +			__entry->comm, __entry->pid,
>> +			__print_symbolic(__entry->policy, SCHEDULING_POLICY),
>> +			__entry->nice, __entry->rt_priority,
>> +			__entry->dl_runtime, __entry->dl_deadline,
>> +			__entry->dl_period)
>> +);
>>  #endif /* _TRACE_SCHED_H */
>>  
>>  /* This part must be outside protection */
>> diff --git a/kernel/fork.c b/kernel/fork.c
>> index 7926993..ac4294a 100644
>> --- a/kernel/fork.c
>> +++ b/kernel/fork.c
>> @@ -1773,6 +1773,7 @@ long _do_fork(unsigned long clone_flags,
>>  		struct pid *pid;
>>  
>>  		trace_sched_process_fork(current, p);
>> +		trace_sched_prio_update(p);
>>  
>>  		pid = get_task_pid(p, PIDTYPE_PID);
>>  		nr = pid_vnr(pid);
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index ce83e39..c729425 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -3708,6 +3708,7 @@ void set_user_nice(struct task_struct *p, long nice)
>>  			resched_curr(rq);
>>  	}
>>  out_unlock:
>> +	trace_sched_prio_update(p);
>>  	task_rq_unlock(rq, p, &rf);
>>  }
>>  EXPORT_SYMBOL(set_user_nice);
>> @@ -3912,6 +3913,8 @@ static void __setscheduler(struct rq *rq, struct
>> task_struct *p,
>>  		p->sched_class = &rt_sched_class;
>>  	else
>>  		p->sched_class = &fair_sched_class;
>> +
>> +	trace_sched_prio_update(p);
>>  }
>>  
> >  static void

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com