[LTP] [PATCH V3 08/23] rt_sigaction/rt_sigprocmark: Replace SA_NOMASK with SA_NODEFER

Cyril Hrubis chrubis@suse.cz
Wed Jul 27 17:35:00 CEST 2016


Hi!
> >> +#define _GNU_SOURCE
> >> #include <stdio.h>
> >> #include <stdlib.h>
> >> #include <string.h>
> >> @@ -409,7 +409,7 @@ void my_signal(int sig, void (*func) ())
> >> 
> >> 	act.sa_handler = func;
> >> 	memset(&act.sa_mask, 0x00, sizeof(sigset_t));
> >> -	act.sa_flags = SA_NOMASK | SA_RESTART;
> >> +	act.sa_flags = SA_NODEFER | SA_RESTART;
> >> 	sigaction(sig, &act, 0);
> >> }
> >> 
> >> diff --git a/testcases/misc/crash/crash02.c b/testcases/misc/crash/crash02.c
> >> index 00fee34..1aba3b9 100644
> >> --- a/testcases/misc/crash/crash02.c
> >> +++ b/testcases/misc/crash/crash02.c
> >> @@ -49,7 +49,6 @@ TODO:
> >> 	* Does a syscall is supposed to send SIGSEGV?
> >> */
> >> 
> >> -#define _GNU_SOURCE
> >> #include <sys/syscall.h>
> >> #include <stdio.h>
> >> #include <stdlib.h>
> >> @@ -358,7 +357,7 @@ void my_signal(int sig, void (*func) ())
> >> 
> >> 	act.sa_handler = func;
> >> 	memset(&act.sa_mask, 0x00, sizeof(sigset_t));
> >> -	act.sa_flags = SA_NOMASK | SA_RESTART;
> >> +	act.sa_flags = SA_NODEFER | SA_RESTART;
> >> 	sigaction(sig, &act, 0);
> >> }
> > 
> > What is the reason to add #define _GNU_SOURCE to crash01 and to remove
> > it from crash02? That looks like unintended change to me.
> 
> The define was needed only for these SA_* symbols which are removed
> so I removed the define as well.

Hmm, so why the patch adds it to crash01.c at all?

-- 
Cyril Hrubis
chrubis@suse.cz


More information about the ltp mailing list