Re: [LTP] [PATCH v4] Refactor pidns30 test using new LTP API

[Richard Palethorpe <rpalethorpe@suse.de>]

Hello,

Cyril Hrubis <chrubis@suse.cz> writes:

> Hi!
>> >> +static void run(void)
>> >>   {
>> >> -	int status;
>> >> -	char buf[5];
>> >> -	pid_t cpid;
>> >> +	const struct tst_clone_args args = { CLONE_NEWPID, SIGCHLD };
>> >>   
>> >> -	setup();
>> >> +	remove_mqueue(mqd);
>> >> +	received = 0;
>> > I wonder if we really need to clear the flag here, as far as I can see
>> > we only modify that variable in the child process and that shouldn't
>> > propagate to parent, unless we pass CLONE_VM of course.
>> Flag reset has been added to be 100% sure we are receiving the signal on 
>> -i > 1.
>
> My point was that the variable is only changed in the child process,
> that means that this change will never make it back to the parent since
> the whole VM is copy-on-write...

Indeed, I shouldn't have merged that.

>
> -- 
> Cyril Hrubis
> chrubis@suse.cz


-- 
Thank you,
Richard.

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp