Re: [PATCH v2 bpf-next 3/5] ftrace: introduce FTRACE_OPS_FL_SHARE_IPMODIFY

[Song Liu <songliubraving@fb.com>]

> On Jul 15, 2022, at 2:29 PM, Steven Rostedt <rostedt@goodmis.org> wrote:
> 
> On Fri, 15 Jul 2022 20:21:49 +0000
> Song Liu <songliubraving@fb.com> wrote:
> 
>>>>> Wouldn't this need to be done anyway if BPF was first and live kernel
>>>>> patching needed the update? An -EAGAIN would not suffice.    
>>>> 
>>>> prepare_direct_functions_for_ipmodify handles BPF-first-livepatch-later
>>>> case. The benefit of prepare_direct_functions_for_ipmodify() is that it 
>>>> holds direct_mutex before ftrace_lock, and keeps holding it if necessary. 
>>>> This is enough to make sure we don't need the wash-rinse-repeat. 
>>>> 
>>>> OTOH, if we wait until __ftrace_hash_update_ipmodify(), we already hold
>>>> ftrace_lock, but not direct_mutex. To make changes to bpf trampoline, we
>>>> have to unlock ftrace_lock and lock direct_mutex to avoid deadlock. 
>>>> However, this means we will need the wash-rinse-repeat.   
>> 
>> What do you think about the prepare_direct_functions_for_ipmodify() 
>> approach? If this is not ideal, maybe we can simplify it so that it only
>> holds direct_mutex (when necessary). The benefit is that we are sure
>> direct_mutex is already held in __ftrace_hash_update_ipmodify(). However, 
>> I think it is not safe to unlock ftrace_lock in __ftrace_hash_update_ipmodify(). 
>> We can get parallel do_for_each_ftrace_rec(), which is dangerous, no? 
> 
> I'm fine with it. But one nit on the logic:
> 
>> int register_ftrace_function(struct ftrace_ops *ops)
>> +	__releases(&direct_mutex)
>> {
>> +	bool direct_mutex_locked;
>> 	int ret;
>> 
>> 	ftrace_ops_init(ops);
>> 
>> +	ret = prepare_direct_functions_for_ipmodify(ops);
>> +	if (ret < 0)
>> +		return ret;
>> +
>> +	direct_mutex_locked = ret == 1;
>> +
> 
> Please make the above:
> 
> 	if (ret < 0)
> 		return ret;
> 	else if (ret == 1)
> 		direct_mutex_locked = true;
> 
> It's much easier to read that way.

Thanks for the clarification! 

Song