Re: [RFC patch 04/19] bpf/tracing: Remove redundant preempt_disable() in __bpf_trace_run()

[Thomas Gleixner]

Steven Rostedt <rostedt@xxxxxxxxxxx> writes:

> On Fri, 14 Feb 2020 14:39:21 +0100
> Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
>
>> __bpf_trace_run() disables preemption around the BPF_PROG_RUN() invocation.
>> 
>> This is redundant because __bpf_trace_run() is invoked from a trace point
>> via __DO_TRACE() which already disables preemption _before_ invoking any of
>> the functions which are attached to a trace point.
>> 
>> Remove it.
>> 
>> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
>> ---
>>  kernel/trace/bpf_trace.c |    2 --
>>  1 file changed, 2 deletions(-)
>> 
>> --- a/kernel/trace/bpf_trace.c
>> +++ b/kernel/trace/bpf_trace.c
>> @@ -1476,9 +1476,7 @@ static __always_inline
>>  void __bpf_trace_run(struct bpf_prog *prog, u64 *args)
>>  {
>
> Should there be a "cant_migrate()" added here?

A cant_sleep() is the right thing to add as this really needs to stay
non-preemptible. Hmm?

>>  	rcu_read_lock();
>> -	preempt_disable();
>>  	(void) BPF_PROG_RUN(prog, args);
>> -	preempt_enable();
>>  	rcu_read_unlock();
>>  }

Thanks,

        tglx