Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [RFC PATCH] trace: prevent preemption in
 perf_ftrace_function_call()
To:     Steven Rostedt <rostedt@goodmis.org>
Cc:     Ingo Molnar <mingo@redhat.com>,
        open list <linux-kernel@vger.kernel.org>
References: <2470f39b-aed1-4e19-9982-206007eb0c6a@linux.alibaba.com>
 <20210923093359.30da8ba6@gandalf.local.home>
From:   =?UTF-8?B?546L6LSH?= <yun.wang@linux.alibaba.com>
Message-ID: <7f4dfb4a-d271-b3c5-f603-06cc789ba9e9@linux.alibaba.com>
Date:   Fri, 24 Sep 2021 10:08:10 +0800
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.13; rv:78.0)
 Gecko/20100101 Thunderbird/78.14.0
MIME-Version: 1.0
In-Reply-To: <20210923093359.30da8ba6@gandalf.local.home>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 2021/9/23 下午9:33, Steven Rostedt wrote:
[snip]
>> ---
>>  kernel/trace/trace_event_perf.c | 14 +++++++++-----
>>  1 file changed, 9 insertions(+), 5 deletions(-)
>>
>> diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c
>> index 6aed10e..5486b18 100644
>> --- a/kernel/trace/trace_event_perf.c
>> +++ b/kernel/trace/trace_event_perf.c
>> @@ -438,15 +438,17 @@ void perf_trace_buf_update(void *record, u16 type)
>>  	int rctx;
>>  	int bit;
>>
>> +	preempt_disable_notrace();
>> +
>>  	if (!rcu_is_watching())
>> -		return;
>> +		goto out;
> 
> You don't need preemption disabled for the above check. It's better to
> disable it here and leave the above return untouched.

I found the rcu tree implementation of rcu_is_watching() will check
this_cpu_ptr(&rcu_data.dynticks), and after that enable the preemption.

If preemption happened after that and before we disable here, there are
still possibility that the CPU changed and make the dynticks checking
invalid, isn't it?

Regards,
Michael Wang

> 
> -- Steve
> 
>>
>>  	if ((unsigned long)ops->private != smp_processor_id())
>> -		return;
>> +		goto out;
>>
>>  	bit = ftrace_test_recursion_trylock(ip, parent_ip);
>>  	if (bit < 0)
>> -		return;
>> +		goto out;
>>
>>  	event = container_of(ops, struct perf_event, ftrace_ops);
>>
>> @@ -468,16 +470,18 @@ void perf_trace_buf_update(void *record, u16 type)
>>
>>  	entry = perf_trace_buf_alloc(ENTRY_SIZE, NULL, &rctx);
>>  	if (!entry)
>> -		goto out;
>> +		goto out_unlock;
>>
>>  	entry->ip = ip;
>>  	entry->parent_ip = parent_ip;
>>  	perf_trace_buf_submit(entry, ENTRY_SIZE, rctx, TRACE_FN,
>>  			      1, &regs, &head, NULL);
>>
>> -out:
>> +out_unlock:
>>  	ftrace_test_recursion_unlock(bit);
>>  #undef ENTRY_SIZE
>> +out:
>> +	preempt_enable_notrace();
>>  }
>>
>>  static int perf_ftrace_function_register(struct perf_event *event)