On Wed, Jun 15, 2016 at 01:20:27PM +0800, Fei Jie wrote:
> * tests/signal_receive.c: New file.
> * tests/signal_receive.test: New test.
> * tests/.gitignore: Add signal_receive.
> * tests/Makefile.am (check_PROGRAMS): Likewise.
> (MISC_TESTS): Add signal_receive.test.
> ---
> tests/.gitignore | 1 +
> tests/Makefile.am | 2 ++
> tests/signal_receive.c | 51
> +++++++++++++++++++++++++++++++++++++++++++++++
> tests/signal_receive.test | 8 ++++++++
> 4 files changed, 62 insertions(+)
> create mode 100644 tests/signal_receive.c
> create mode 100755 tests/signal_receive.test
>
> diff --git a/tests/.gitignore b/tests/.gitignore
> index 67d0d66..72e93d8 100644
> --- a/tests/.gitignore
> +++ b/tests/.gitignore
> @@ -231,6 +231,7 @@ shmxt
> sigaction
> sigaltstack
> siginfo
> +signal_receive
> signalfd
> sigreturn
> sleep
> diff --git a/tests/Makefile.am b/tests/Makefile.am
> index a136e2a..94d3d23 100644
> --- a/tests/Makefile.am
> +++ b/tests/Makefile.am
> @@ -285,6 +285,7 @@ check_PROGRAMS = \
> sigaction \
> sigaltstack \
> siginfo \
> + signal_receive \
> signalfd \
> sigreturn \
> sleep \
> @@ -662,6 +663,7 @@ MISC_TESTS = \
> qual_syscall.test \
> redirect.test \
> restart_syscall.test \
> + signal_receive.test \
> strace-E.test \
> strace-S.test \
> strace-T.test \
> diff --git a/tests/signal_receive.c b/tests/signal_receive.c
> new file mode 100644
> index 0000000..c3027a4
> --- /dev/null
> +++ b/tests/signal_receive.c
> @@ -0,0 +1,51 @@
> +#include "tests.h"
> +#include <assert.h>
> +#include <signal.h>
> +#include <stdio.h>
> +#include <unistd.h>
> +
> +void sig_print(const char *signame)
> +{
> + printf("kill\\(%d, %s\\) = 0\n"
> + "--- %s {si_signo=%s, si_code=SI_USER, si_pid=%d,"
> + " si_uid=%d} ---\n",
> + getpid(), signame, signame, signame, getpid(), getuid());
Invoking getpid() and getuid() so many times looks odd.
Let's invoke them once and pass the result to this function.
> +}
> +
> +static void
> +handler(int sig)
> +{
> +}
> +
> +int
> +main(void)
> +{
> + int sig;
> + const struct sigaction act = { .sa_handler = handler };
> + for (sig = 1; sig <= 31; sig++)
> + {
> + if( sig != SIGKILL && sig != SIGSTOP)
> + {
> + sigaction(sig, &act, NULL);
> + }
> + }
Opening braces in this function are not in the right place.
The tradition is to put them on the same line with the statement.
> + sigset_t mask;
> + sigemptyset(&mask);
> + for (sig = 1; sig <= 31; sig++)
> + {
> + sigaddset(&mask, sig);
> + }
This loop could be merged with the previous one.
> + sigprocmask(SIG_UNBLOCK, &mask, NULL);
> +
> + for (sig = 1; sig <= 31; sig++)
> + {
> + if(sig != SIGKILL && sig != SIGSTOP)
> + {
> + assert(kill(getpid(), sig) == 0);
I'd be very much surprised if this kill failed, but if it did,
I'd like to see the reason. Let's use perror_msg_and_fail
in cases like this.
> + sig_print(signal2name(sig));
> + }
> + }
> +
> + return 0;
> +}
> diff --git a/tests/signal_receive.test b/tests/signal_receive.test
> new file mode 100755
> index 0000000..2e3e98d
> --- /dev/null
> +++ b/tests/signal_receive.test
> @@ -0,0 +1,8 @@
> +#!/bin/sh
> +
> +. "${srcdir=.}/init.sh"
A comment explaining what the test does would be a good addition.
> +run_prog > /dev/null
> +run_strace -a16 -esignal $args > "$EXP"
> +match_grep "$LOG" "$EXP"
> +rm -f "$EXP"
Is there any particular reason to use less precise match_grep
rather than run_strace_match_diff?
--
ldv
pgpCDI5MqHo9t.pgp
Description: PGP signature
------------------------------------------------------------------------------ What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. http://pubads.g.doubleclick.net/gampad/clk?id=1444514421&iu=/41014381
_______________________________________________ Strace-devel mailing list Strace-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/strace-devel
