Re: [PATCH 1/2 v5 typo updated] tracing/uprobes: Support ftrace_event_file base multibuffer

["zhangwei(Jovi)" <jovi.zhangwei@huawei.com>]

On 2013/7/4 15:41, Namhyung Kim wrote:
> Hi Jovi,
> 
> Just a few of dummy questions..
> 
> 
> On Thu, 4 Jul 2013 15:01:10 +0800, zhangwei wrote:
>> Support multi-buffer on uprobe-based dynamic events by
>> using ftrace_event_file.
>>
>> This patch is based kprobe-based dynamic events multibuffer
>> support work initially, commited by Masami(commit 41a7dd420c),
>> but revised as below:
>>
>> Oleg changed the kprobe-based multibuffer design from
>> array-pointers of ftrace_event_file into simple list,
>> so this patch also change to the list design.
>>
>> rcu_read_lock/unlock added into uprobe_trace_func/uretprobe_trace_func,
>> to synchronize with ftrace_event_file list add and delete.
>>
>> Even though we allow multi-uprobes instances now,
>> but TP_FLAG_PROFILE/TP_FLAG_TRACE are still mutually exclusive
>> in probe_event_enable currently, this means we cannot allow
>> one user is using uprobe-tracer, and another user is using
>> perf-probe on same uprobe concurrently.
>> (Perhaps this will be fix in future, kprobe dont't have this
>> limitation now)
> 
> So why does this limitation exist?  Didn't we support this kind of thing
> in the original code?
> 
Yes, it existed(maybe not exist before uprobe pre-filter work), because uprobe filter
is associated with trace_uprobe tightly at present, so we cannot assign
TP_FLAG_PROFILE/TP_FLAG_TRACE for same trace_uprobe with different filter.

Perhaps we need to remove the limitation in future.

>>
>> Signed-off-by: zhangwei(Jovi) <jovi.zhangwei@huawei.com>
>> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
>> Cc: Frederic Weisbecker <fweisbec@gmail.com>
>> Cc: Oleg Nesterov <oleg@redhat.com>
>> Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
>> ---
> [SNIP]
>> +	rcu_read_lock();
>> +	list_for_each_entry(link, &tu->files, list)
> 
> list_for_each_entry_rcu() ?
> 
I haven't noticed this, thanks, I will update it.
> 
>> +		uprobe_trace_print(tu, 0, regs, link->file);
>> +	rcu_read_unlock();
>> +
>>  	return 0;
>>  }
>>
>>  static void uretprobe_trace_func(struct trace_uprobe *tu, unsigned long func,
>>  				struct pt_regs *regs)
>>  {
>> -	uprobe_trace_print(tu, func, regs);
>> +	struct event_file_link *link;
>> +
>> +	rcu_read_lock();
>> +	list_for_each_entry(link, &tu->files, list)
> 
> Ditto.
> 
> 
>> +		uprobe_trace_print(tu, func, regs, link->file);
>> +	rcu_read_unlock();
>>  }
> [SNIP]
>> -static void probe_event_disable(struct trace_uprobe *tu, int flag)
>> +static struct event_file_link *
>> +find_event_file_link(struct trace_uprobe *tu, struct ftrace_event_file *file)
>> +{
>> +	struct event_file_link *link;
>> +
>> +	list_for_each_entry(link, &tu->files, list)
> 
> Not sure of this case. ;)
> 
Yes, _rcu is not needed in here, it's only called in event disable serialized case.

> Thanks,
> Namhyung
> 
>> +		if (link->file == file)
>> +			return link;
>> +
>> +	return NULL;
>> +}
>> +
>> +static void
>> +probe_event_disable(struct trace_uprobe *tu, struct ftrace_event_file *file)
>>  {
>>  	if (!is_trace_uprobe_enabled(tu))
>>  		return;
>>
>> +	if (file) {
>> +		struct event_file_link *link;
>> +
>> +		link = find_event_file_link(tu, file);
>> +		if (!link)
>> +			return;
>> +
>> +		list_del_rcu(&link->list);
>> +		/* synchronize with uprobe_trace_func/uretprobe_trace_func */
>> +		synchronize_sched();
>> +		kfree(link);
>> +
>> +		if (!list_empty(&tu->files))
>> +			return;
>> +	}
>> +
>>  	WARN_ON(!uprobe_filter_is_empty(&tu->filter));
>>
>>  	uprobe_unregister(tu->inode, tu->offset, &tu->consumer);
>> -	tu->flags &= ~flag;
>> +	tu->flags &= file ? ~TP_FLAG_TRACE : ~TP_FLAG_PROFILE;
>>  }
>>
>>  static int uprobe_event_define_fields(struct ftrace_event_call *event_call)
>> @@ -867,21 +947,22 @@ static
>>  int trace_uprobe_register(struct ftrace_event_call *event, enum trace_reg type, void *data)
>>  {
>>  	struct trace_uprobe *tu = event->data;
>> +	struct ftrace_event_file *file = data;
>>
>>  	switch (type) {
>>  	case TRACE_REG_REGISTER:
>> -		return probe_event_enable(tu, TP_FLAG_TRACE, NULL);
>> +		return probe_event_enable(tu, file, NULL);
>>
>>  	case TRACE_REG_UNREGISTER:
>> -		probe_event_disable(tu, TP_FLAG_TRACE);
>> +		probe_event_disable(tu, file);
>>  		return 0;
>>
>>  #ifdef CONFIG_PERF_EVENTS
>>  	case TRACE_REG_PERF_REGISTER:
>> -		return probe_event_enable(tu, TP_FLAG_PROFILE, uprobe_perf_filter);
>> +		return probe_event_enable(tu, NULL, uprobe_perf_filter);
>>
>>  	case TRACE_REG_PERF_UNREGISTER:
>> -		probe_event_disable(tu, TP_FLAG_PROFILE);
>> +		probe_event_disable(tu, NULL);
>>  		return 0;
>>
>>  	case TRACE_REG_PERF_OPEN:
>> -- 1.7.9.7
> 
> .
>