[LTP] [PATCH V2 09/11] syscalls/waitpid: adapt reap_children() to test stopped children

[Stanislav Kholmanskikh]

On 08/25/2016 02:39 PM, Cyril Hrubis wrote:
> Hi!
>> If the pid returned by waitpid() is the pid of a stopped child,
>> we send it the SIGCONT signal.
>>
>> Signed-off-by: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com>
>> ---
>> This is a new patch in the series, extracted from V1 of waitpid13.c
>>
>>  testcases/kernel/syscalls/waitpid/waitpid_common.h |   17 +++++++++++++++++
>>  1 files changed, 17 insertions(+), 0 deletions(-)
>>
>> diff --git a/testcases/kernel/syscalls/waitpid/waitpid_common.h b/testcases/kernel/syscalls/waitpid/waitpid_common.h
>> index f724a17..7b67a06 100644
>> --- a/testcases/kernel/syscalls/waitpid/waitpid_common.h
>> +++ b/testcases/kernel/syscalls/waitpid/waitpid_common.h
>> @@ -131,6 +131,23 @@ static int reap_children(pid_t wp_pid, int wp_opts, pid_t *children, int len)
>>  			return -1;
>>  		}
>>  
>> +		if (WIFSTOPPED(status)) {
>> +			if (WSTOPSIG(status) != SIGSTOP) {
>> +				tst_res(TFAIL,
>> +					"Pid %d: expected SIGSTOP, got %d",
>> +					pid, WSTOPSIG(status));
>> +				return -1;
>> +			}
> 
> Maybe we should print tst_res(TINFO, "Seding SIGCONT to %i", pid); here.

Ok.

> 
>> +			if (kill(pid, SIGCONT) < 0) {
>> +				tst_res(TFAIL | TERRNO,
>> +					"kill(%d, SIGCONT) failed", pid);
>> +				return -1;
>> +			}
>> +
>> +			continue;
>> +		}
>> +
>>  		for (i = 0; i < len; i++) {
>>  			if (pid == children[i]) {
>>  				children[i] = 0;
> 
> Otherwise it looks good.
> 
> I guess that there is no real chance that some of the children in
> testcases that use this function would be SIGSTOPed accidentally and
> this code will make that bug disappear...
> 

Yes. I agree.