Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH 5/5] tracing: New functions for kernel access to Ftrace
 instances
To:     Steven Rostedt <rostedt@goodmis.org>
Cc:     linux-kernel@vger.kernel.org, Joe Jin <joe.jin@oracle.com>,
        Srinivas Eeda <srinivas.eeda@oracle.com>,
        Aruna Ramakrishna <aruna.ramakrishna@oracle.com>
References: <1565805327-579-1-git-send-email-divya.indi@oracle.com>
 <1565805327-579-6-git-send-email-divya.indi@oracle.com>
 <20191015191917.6287ba10@gandalf.local.home>
From:   Divya Indi <divya.indi@oracle.com>
Message-ID: <054da671-d159-080f-04c8-0b1d34cf037d@oracle.com>
Date:   Wed, 16 Oct 2019 16:53:04 -0700
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.14; rv:60.0)
 Gecko/20100101 Thunderbird/60.9.0
MIME-Version: 1.0
In-Reply-To: <20191015191917.6287ba10@gandalf.local.home>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Content-Language: en-US
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Steven,

Please find my comments inline -

On 10/15/19 4:19 PM, Steven Rostedt wrote:
> On Wed, 14 Aug 2019 10:55:27 -0700
> Divya Indi <divya.indi@oracle.com> wrote:
>
>> Adding 2 new functions -
>> 1) trace_array_lookup : Look up and return a trace array, given its
>> name.
>> 2) trace_array_set_clr_event : Enable/disable event recording to the
>> given trace array.
>>
>> NOTE: trace_array_lookup returns a trace array and also increments the
>> reference counter associated with the returned trace array. Make sure to
>> call trace_array_put() once the use is done so that the instance can be
>> removed at a later time.
>>
>> Example use:
>>
>> tr = trace_array_lookup("foo-bar");
> Should probably be renamed to: trace_array_get_by_name("foo-bar")
>
> As the name should show that it adds a ref count.

Makes sense! Will make this change.

>
> But if we make the change I suggested before, where we do not need to
> do the put before calling the destroy, we should have:
>
>
>> if (!tr)
> {
>> 	tr = trace_array_create("foo-bar");
> 	trace_array_get(tr);

This would need a corresponding trace_array_put(?).
Additionally, this would again traverse the list of instances, So I
went with incrementing the ref ctr inside the function.

> }
>
>
>> // Log to tr
>> // Enable/disable events to tr
>> trace_array_set_clr_event(tr, _THIS_IP,"system","event",1);
> What's with the __THIS_IP?

My bad! Maybe remnants from a trace_array_printk copy-paste.
Will correct this.

>
>
>> // Done using tr
>> trace_array_put(tr);
>> ..
>>
>> Also, use trace_array_set_clr_event to enable/disable events to a trace array.
>> So now we no longer need to have ftrace_set_clr_event as an exported
>> API.
>>
>> Signed-off-by: Divya Indi <divya.indi@oracle.com>
>> ---
>>   include/linux/trace.h        |  2 ++
>>   include/linux/trace_events.h |  3 ++-
>>   kernel/trace/trace.c         | 28 ++++++++++++++++++++++++++++
>>   kernel/trace/trace_events.c  | 23 ++++++++++++++++++++++-
>>   4 files changed, 54 insertions(+), 2 deletions(-)
>>
>> diff --git a/include/linux/trace.h b/include/linux/trace.h
>> index 2c782d5..05164bb 100644
>> --- a/include/linux/trace.h
>> +++ b/include/linux/trace.h
>> @@ -32,6 +32,8 @@ int trace_array_printk(struct trace_array *tr, unsigned long ip,
>>   struct trace_array *trace_array_create(const char *name);
>>   int trace_array_destroy(struct trace_array *tr);
>>   void trace_array_put(struct trace_array *tr);
>> +struct trace_array *trace_array_lookup(const char *name);
>> +
>>   #endif	/* CONFIG_TRACING */
>>   
>>   #endif	/* _LINUX_TRACE_H */
>> diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
>> index 8a62731..05a7514 100644
>> --- a/include/linux/trace_events.h
>> +++ b/include/linux/trace_events.h
>> @@ -540,7 +540,8 @@ extern int trace_define_field(struct trace_event_call *call, const char *type,
>>   #define is_signed_type(type)	(((type)(-1)) < (type)1)
>>   
>>   int trace_set_clr_event(const char *system, const char *event, int set);
>> -
>> +int trace_array_set_clr_event(struct trace_array *tr, const char *system,
>> +		const char *event, int set);
>>   /*
>>    * The double __builtin_constant_p is because gcc will give us an error
>>    * if we try to allocate the static variable to fmt if it is not a
>> diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
>> index 7b6a37a..e394d55 100644
>> --- a/kernel/trace/trace.c
>> +++ b/kernel/trace/trace.c
>> @@ -8514,6 +8514,34 @@ static int instance_rmdir(const char *name)
>>   	return ret;
>>   }
>>   
>> +/**
>> + * trace_array_lookup - Lookup the trace array, given its name.
>> + * @name: The name of the trace array to be looked up.
>> + *
>> + * Lookup and return the trace array associated with @name.
>> + *
>> + * NOTE: The reference counter associated with the returned trace array is
>> + * incremented to ensure it is not freed when in use. Make sure to call
>> + * "trace_array_put" for this trace array when its use is done.
>> + */
>> +struct trace_array *trace_array_lookup(const char *name)
>> +{
>> +	struct trace_array *tr;
>> +
>> +	mutex_lock(&trace_types_lock);
>> +
>> +	list_for_each_entry(tr, &ftrace_trace_arrays, list) {
>> +		if (tr->name && strcmp(tr->name, name) == 0) {
>> +			tr->ref++;
>> +			mutex_unlock(&trace_types_lock);
>> +			return tr;
>> +		}
>> +	}
>> +	mutex_unlock(&trace_types_lock);
>> +	return NULL;
>> +}
>> +EXPORT_SYMBOL_GPL(trace_array_lookup);
>> +
>>   static __init void create_trace_instances(struct dentry *d_tracer)
>>   {
>>   	trace_instance_dir = tracefs_create_instance_dir("instances", d_tracer,
>> diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
>> index 2621995..96dd997 100644
>> --- a/kernel/trace/trace_events.c
>> +++ b/kernel/trace/trace_events.c
>> @@ -834,7 +834,6 @@ static int ftrace_set_clr_event(struct trace_array *tr, char *buf, int set)
>>   
>>   	return ret;
>>   }
>> -EXPORT_SYMBOL_GPL(ftrace_set_clr_event);
>>   
>>   /**
>>    * trace_set_clr_event - enable or disable an event
>> @@ -859,6 +858,28 @@ int trace_set_clr_event(const char *system, const char *event, int set)
>>   }
>>   EXPORT_SYMBOL_GPL(trace_set_clr_event);
>>   
>> +/**
>> + * trace_array_set_clr_event - enable or disable an event for a trace array
>> + * @system: system name to match (NULL for any system)
>> + * @event: event name to match (NULL for all events, within system)
>> + * @set: 1 to enable, 0 to disable
> We probably should make this a boolean.

This would internally call ftrace_set_clr_event which would expect 0/1.
But I agree making it boolean would be more intuitive.

Thanks,
Divya

>
> -- Steve
>
>> + *
>> + * This is a way for other parts of the kernel to enable or disable
>> + * event recording to instances.
>> + *
>> + * Returns 0 on success, -EINVAL if the parameters do not match any
>> + * registered events.
>> + */
>> +int trace_array_set_clr_event(struct trace_array *tr, const char *system,
>> +		const char *event, int set)
>> +{
>> +	if (!tr)
>> +		return -ENOENT;
>> +
>> +	return __ftrace_set_clr_event(tr, NULL, system, event, set);
>> +}
>> +EXPORT_SYMBOL_GPL(trace_array_set_clr_event);
>> +
>>   /* 128 should be much more than enough */
>>   #define EVENT_BUF_SIZE		127
>>