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 v2 2/2] net: add trace events for net_device refcnt
To:     Tony Lu <tonylu@linux.alibaba.com>, davem@davemloft.net
Cc:     xiyou.wangcong@gmail.com, eric.dumazet@gmail.com,
        shemminger@osdl.org, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
References: <20191112130510.91570-1-tonylu@linux.alibaba.com>
 <20191112130510.91570-2-tonylu@linux.alibaba.com>
From:   David Ahern <dsahern@gmail.com>
Message-ID: <ea3bf073-68af-d899-2664-fef54953a68d@gmail.com>
Date:   Tue, 12 Nov 2019 08:27:31 -0700
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.14; rv:68.0)
 Gecko/20100101 Thunderbird/68.2.2
MIME-Version: 1.0
In-Reply-To: <20191112130510.91570-2-tonylu@linux.alibaba.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 11/12/19 6:05 AM, Tony Lu wrote:
> diff --git a/include/trace/events/net.h b/include/trace/events/net.h
> index 3b28843652d2..3bf6dd738882 100644
> --- a/include/trace/events/net.h
> +++ b/include/trace/events/net.h
> @@ -326,6 +326,47 @@ DEFINE_EVENT(net_dev_rx_exit_template, netif_receive_skb_list_exit,
>  	TP_ARGS(ret)
>  );
>  
> +DECLARE_EVENT_CLASS(net_dev_refcnt_template,
> +
> +	TP_PROTO(struct net_device *dev, void *location),
> +
> +	TP_ARGS(dev, location),
> +
> +	TP_STRUCT__entry(
> +		__string(	name,		dev->name	)
> +		__field(	int,		refcnt		)
> +		__field(	void *,		location	)
> +	),
> +
> +	TP_fast_assign(
> +		int i, refcnt = 0;
> +
> +		for_each_possible_cpu(i)
> +			refcnt += *per_cpu_ptr(dev->pcpu_refcnt, i);

Rather than copying the definition of netdev_refcnt_read here, so just
use it.

> +
> +		__assign_str(name, dev->name);
> +		__entry->refcnt = refcnt;
> +		__entry->location = location;
> +	),
> +
> +	TP_printk("dev=%s refcnt=%d location=%p",
> +		__get_str(name), __entry->refcnt, __entry->location)
> +);
> +
> +DEFINE_EVENT(net_dev_refcnt_template, net_dev_put,
> +
> +	TP_PROTO(struct net_device *dev, void *location),
> +
> +	TP_ARGS(dev, location)
> +);
> +
> +DEFINE_EVENT(net_dev_refcnt_template, net_dev_hold,
> +
> +	TP_PROTO(struct net_device *dev, void *location),
> +
> +	TP_ARGS(dev, location)
> +);
> +
>  #endif /* _TRACE_NET_H */
>  

The location alone does nothing for resolving reference count leaks; you
really have to use stack traces to pair up the hold and put and to give
context for what are repetitive locations for the hold and put.