public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
* [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison
@ 2020-04-10 18:55 Alistair Francis
  2020-04-10 18:55 ` [PATCH v3 2/2] linux: Add wait4 test case Alistair Francis
                   ` (2 more replies)
  0 siblings, 3 replies; 5+ messages in thread
From: Alistair Francis @ 2020-04-10 18:55 UTC (permalink / raw)
  To: libc-alpha; +Cc: alistair23, Alistair Francis

Patch 600f00b "linux: Use long time_t for wait4/getrusage" introduced
two bugs:
 - The usage32 struct was set if the wait4 syscall had an error.
 - For 32-bit systems the usage struct was set even if it was specified
   as NULL.

This patch fixes the two issues.
---
 sysdeps/unix/sysv/linux/wait4.c | 19 +++++++++----------
 1 file changed, 9 insertions(+), 10 deletions(-)

diff --git a/sysdeps/unix/sysv/linux/wait4.c b/sysdeps/unix/sysv/linux/wait4.c
index d14bd4da27..44c55f60fe 100644
--- a/sysdeps/unix/sysv/linux/wait4.c
+++ b/sysdeps/unix/sysv/linux/wait4.c
@@ -29,13 +29,13 @@ __wait4_time64 (pid_t pid, int *stat_loc, int options, struct __rusage64 *usage)
 # if __KERNEL_OLD_TIMEVAL_MATCHES_TIMEVAL64
   return SYSCALL_CANCEL (wait4, pid, stat_loc, options, usage);
 # else
+  pid_t ret;
   struct __rusage32 usage32;
-  pid_t ret = SYSCALL_CANCEL (wait4, pid, stat_loc, options, &usage32);
 
-  if (ret != 0)
-    return ret;
+  ret = SYSCALL_CANCEL (wait4, pid, stat_loc, options,
+                        usage != NULL ? &usage32 : NULL);
 
-  if (usage != NULL)
+  if (ret > 0 && usage != NULL)
     rusage32_to_rusage64 (&usage32, usage);
 
   return ret;
@@ -114,15 +114,14 @@ libc_hidden_def (__wait4_time64)
 pid_t
 __wait4 (pid_t pid, int *stat_loc, int options, struct rusage *usage)
 {
-  pid_t ret ;
+  pid_t ret;
   struct __rusage64 usage64;
 
-  ret = __wait4_time64 (pid, stat_loc, options, &usage64);
+  ret = __wait4_time64 (pid, stat_loc, options,
+                        usage != NULL ? &usage64 : NULL);
 
-  if (ret != 0)
-    return ret;
-
-  rusage64_to_rusage (&usage64, usage);
+  if (ret > 0 && usage != 0)
+    rusage64_to_rusage (&usage64, usage);
 
   return ret;
 }
-- 
2.26.0


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH v3 2/2] linux: Add wait4 test case
  2020-04-10 18:55 [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison Alistair Francis
@ 2020-04-10 18:55 ` Alistair Francis
  2020-04-10 18:55 ` [PATCH v4 2/2] posix: " Alistair Francis
  2020-04-14 12:07 ` [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison Andreas Schwab
  2 siblings, 0 replies; 5+ messages in thread
From: Alistair Francis @ 2020-04-10 18:55 UTC (permalink / raw)
  To: libc-alpha; +Cc: alistair23, Alistair Francis

Add a simple test that checks if the wait4 syscall works without an
rusage argument. This test would have caught the bugs introduced in
commit 600f00b "linux: Use long time_t for wait4/getrusage".
---
 sysdeps/unix/sysv/linux/Makefile    |  2 +-
 sysdeps/unix/sysv/linux/tst-wait4.c | 51 +++++++++++++++++++++++++++++
 2 files changed, 52 insertions(+), 1 deletion(-)
 create mode 100644 sysdeps/unix/sysv/linux/tst-wait4.c

diff --git a/sysdeps/unix/sysv/linux/Makefile b/sysdeps/unix/sysv/linux/Makefile
index 60dc5cf9e5..d5fdb4369a 100644
--- a/sysdeps/unix/sysv/linux/Makefile
+++ b/sysdeps/unix/sysv/linux/Makefile
@@ -98,7 +98,7 @@ tests += tst-clone tst-clone2 tst-clone3 tst-fanotify tst-personality \
 	 tst-quota tst-sync_file_range tst-sysconf-iov_max tst-ttyname \
 	 test-errno-linux tst-memfd_create tst-mlock2 tst-pkey \
 	 tst-rlimit-infinity tst-ofdlocks tst-gettid tst-gettid-kill \
-	 tst-tgkill
+	 tst-tgkill tst-wait4
 tests-internal += tst-ofdlocks-compat tst-sigcontext-get_pc
 
 CFLAGS-tst-sigcontext-get_pc.c = -fasynchronous-unwind-tables
diff --git a/sysdeps/unix/sysv/linux/tst-wait4.c b/sysdeps/unix/sysv/linux/tst-wait4.c
new file mode 100644
index 0000000000..44287797a7
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/tst-wait4.c
@@ -0,0 +1,51 @@
+/* Smoke test for the wait4 system call.
+   Copyright (C) 2019-2020 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <support/check.h>
+#include <support/namespace.h>
+#include <sys/wait.h>
+#include <sys/resource.h>
+#include <sys/types.h>
+#include <unistd.h>
+#include <stdlib.h>
+
+/* Check that wait4 works with no rusage arg.  */
+static void
+test_wait4_no_usage (void)
+{
+  pid_t child = fork ();
+
+  if (child == 0)
+    {
+      pause ();
+      exit (0);
+    }
+
+  wait4 (child, &(int){0}, WNOHANG, NULL);
+  kill (child, SIGKILL);
+}
+
+static int
+do_test (void)
+{
+  test_wait4_no_usage();
+
+  return 0;
+}
+
+#include <support/test-driver.c>
-- 
2.26.0


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH v4 2/2] posix: Add wait4 test case
  2020-04-10 18:55 [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison Alistair Francis
  2020-04-10 18:55 ` [PATCH v3 2/2] linux: Add wait4 test case Alistair Francis
@ 2020-04-10 18:55 ` Alistair Francis
  2020-04-14 12:08   ` Andreas Schwab
  2020-04-14 12:07 ` [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison Andreas Schwab
  2 siblings, 1 reply; 5+ messages in thread
From: Alistair Francis @ 2020-04-10 18:55 UTC (permalink / raw)
  To: libc-alpha; +Cc: alistair23, Alistair Francis

Add a wait4 test case based on the waitid test case.
---
 posix/Makefile    |   3 +-
 posix/tst-wait4.c | 209 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 211 insertions(+), 1 deletion(-)
 create mode 100644 posix/tst-wait4.c

diff --git a/posix/Makefile b/posix/Makefile
index 4507d84bf1..605ddbade8 100644
--- a/posix/Makefile
+++ b/posix/Makefile
@@ -82,7 +82,8 @@ tests		:= test-errno tstgetopt testfnm runtests runptests \
 		   bug-regex29 bug-regex30 bug-regex31 bug-regex32 \
 		   tst-nice tst-nanosleep tst-regex2 \
 		   transbug tst-rxspencer tst-pcre tst-boost \
-		   bug-ga1 tst-vfork1 tst-vfork2 tst-vfork3 tst-waitid \
+		   bug-ga1 tst-vfork1 tst-vfork2 tst-vfork3 \
+		   tst-waitid tst-wait4 \
 		   tst-getaddrinfo2 bug-glob2 bug-glob3 tst-sysconf \
 		   tst-execvp1 tst-execvp2 tst-execlp1 tst-execlp2 \
 		   tst-execv1 tst-execv2 tst-execl1 tst-execl2 \
diff --git a/posix/tst-wait4.c b/posix/tst-wait4.c
new file mode 100644
index 0000000000..5696a3456c
--- /dev/null
+++ b/posix/tst-wait4.c
@@ -0,0 +1,209 @@
+/* Tests for wait4.
+   Copyright (C) 2020 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <sys/wait.h>
+#include <sys/resource.h>
+#include <signal.h>
+#include <time.h>
+#include <stdatomic.h>
+#include <stdbool.h>
+
+#include <support/xsignal.h>
+#include <support/xunistd.h>
+#include <support/check.h>
+#include <support/process_state.h>
+
+static void
+test_child (void)
+{
+  /* First thing, we stop ourselves.  */
+  raise (SIGSTOP);
+
+  /* Hey, we got continued!  */
+  while (1)
+    pause ();
+}
+
+#ifndef WEXITED
+# define WEXITED        0
+# define WCONTINUED     0
+# define WSTOPPED       WUNTRACED
+#endif
+
+/* Set with only SIGCHLD on do_test_waitid.  */
+static sigset_t chldset;
+
+#ifdef SA_SIGINFO
+static void
+sigchld (int signo, siginfo_t *info, void *ctx)
+{
+}
+#endif
+
+static void
+check_sigchld (int code, int status, pid_t pid)
+{
+#ifdef SA_SIGINFO
+  siginfo_t siginfo;
+  TEST_COMPARE (sigwaitinfo (&chldset, &siginfo), SIGCHLD);
+
+  TEST_COMPARE (siginfo.si_signo, SIGCHLD);
+  TEST_COMPARE (siginfo.si_code, code);
+  TEST_COMPARE (siginfo.si_status, status);
+  TEST_COMPARE (siginfo.si_pid, pid);
+#endif
+}
+
+static int
+do_test_wait4 (pid_t pid)
+{
+  /* Adding process_state_tracing_stop ('t') allows the test to work under
+     trace programs such as ptrace.  */
+  enum support_process_state stop_state = support_process_state_stopped
+                                          | support_process_state_tracing_stop;
+
+  support_process_state_wait (pid, stop_state);
+
+  check_sigchld (CLD_STOPPED, SIGSTOP, pid);
+
+  pid_t ret;
+  int wstatus;
+  struct rusage rusage;
+
+  ret = wait4 (pid, &wstatus, WUNTRACED|WCONTINUED|WNOHANG, NULL);
+  if (ret == -1 && errno == ENOTSUP)
+    FAIL_RET ("waitid WNOHANG on stopped: %m");
+  TEST_COMPARE (ret, pid);
+  TEST_VERIFY (WIFSTOPPED (wstatus));
+
+  /* Issue again but with struct rusage input.  */
+  ret = wait4 (pid, &wstatus, WUNTRACED|WCONTINUED|WNOHANG, &rusage);
+  /* With WNOHANG and WUNTRACED, if the children has not changes its state
+     since previous call the expected result it 0.  */
+  TEST_COMPARE (ret, 0);
+
+  /* Some sanity tests to check if 'wtatus' and 'rusage' possible
+     input values.  */
+  ret = wait4 (pid, NULL, WUNTRACED|WCONTINUED|WNOHANG, &rusage);
+  TEST_COMPARE (ret, 0);
+  ret = wait4 (pid, NULL, WUNTRACED|WCONTINUED|WNOHANG, NULL);
+  TEST_COMPARE (ret, 0);
+
+  if (kill (pid, SIGCONT) != 0)
+    FAIL_RET ("kill (%d, SIGCONT): %m\n", pid);
+
+  /* Wait for the child to have continued.  */
+  support_process_state_wait (pid, support_process_state_sleeping);
+
+#if WCONTINUED != 0
+  check_sigchld (CLD_CONTINUED, SIGCONT, pid);
+
+  ret = wait4 (pid, &wstatus, WCONTINUED|WNOHANG, NULL);
+  TEST_COMPARE (ret, pid);
+  TEST_VERIFY (WIFCONTINUED (wstatus));
+
+  /* Issue again but with struct rusage input.  */
+  ret = wait4 (pid, &wstatus, WUNTRACED|WCONTINUED|WNOHANG, &rusage);
+  /* With WNOHANG and WUNTRACED, if the children has not changes its state
+     since previous call the expected result it 0.  */
+  TEST_COMPARE (ret, 0);
+
+  /* Now stop him again and test waitpid with WCONTINUED.  */
+  if (kill (pid, SIGSTOP) != 0)
+    FAIL_RET ("kill (%d, SIGSTOP): %m\n", pid);
+
+  /* Wait the child stop.  The waitid call below will block until it has
+     stopped, but if we are real quick and enter the waitid system call
+     before the SIGCHLD has been generated, then it will be discarded and
+     never delivered.  */
+  support_process_state_wait (pid, stop_state);
+
+  ret = wait4 (pid, &wstatus, WUNTRACED|WNOHANG, &rusage);
+  TEST_COMPARE (ret, pid);
+
+  check_sigchld (CLD_STOPPED, SIGSTOP, pid);
+
+  if (kill (pid, SIGCONT) != 0)
+    FAIL_RET ("kill (%d, SIGCONT): %m\n", pid);
+
+  /* Wait for the child to have continued.  */
+  support_process_state_wait (pid, support_process_state_sleeping);
+
+  check_sigchld (CLD_CONTINUED, SIGCONT, pid);
+
+  ret = wait4 (pid, &wstatus, WCONTINUED|WNOHANG, NULL);
+  TEST_COMPARE (ret, pid);
+  TEST_VERIFY (WIFCONTINUED (wstatus));
+#endif
+
+  /* Die, child, die!  */
+  if (kill (pid, SIGKILL) != 0)
+    FAIL_RET ("kill (%d, SIGKILL): %m\n", pid);
+
+  support_process_state_wait (pid, support_process_state_zombie);
+
+  ret = wait4 (pid, &wstatus, 0, &rusage);
+  TEST_COMPARE (ret, pid);
+  TEST_VERIFY (WIFSIGNALED (wstatus));
+  TEST_VERIFY (WTERMSIG (wstatus) == SIGKILL);
+
+  check_sigchld (CLD_KILLED, SIGKILL, pid);
+
+  return 0;
+}
+
+static int
+do_test (void)
+{
+#ifdef SA_SIGINFO
+  {
+    struct sigaction sa;
+    sa.sa_flags = SA_SIGINFO | SA_RESTART;
+    sa.sa_sigaction = sigchld;
+    sigemptyset (&sa.sa_mask);
+    xsigaction (SIGCHLD, &sa, NULL);
+  }
+#endif
+
+  sigemptyset (&chldset);
+  sigaddset (&chldset, SIGCHLD);
+
+  /* The SIGCHLD shall has blocked at the time of the call to sigwait;
+     otherwise, the behavior is undefined.  */
+  sigprocmask (SIG_BLOCK, &chldset, NULL);
+
+  pid_t pid = xfork ();
+  if (pid == 0)
+    {
+      test_child ();
+      _exit (127);
+    }
+
+  do_test_wait4 (pid);
+
+  xsignal (SIGCHLD, SIG_IGN);
+  kill (pid, SIGKILL);          /* Make sure it's dead if we bailed early.  */
+
+  return 0;
+}
+
+#include <support/test-driver.c>
-- 
2.26.0


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison
  2020-04-10 18:55 [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison Alistair Francis
  2020-04-10 18:55 ` [PATCH v3 2/2] linux: Add wait4 test case Alistair Francis
  2020-04-10 18:55 ` [PATCH v4 2/2] posix: " Alistair Francis
@ 2020-04-14 12:07 ` Andreas Schwab
  2 siblings, 0 replies; 5+ messages in thread
From: Andreas Schwab @ 2020-04-14 12:07 UTC (permalink / raw)
  To: Alistair Francis via Libc-alpha; +Cc: Alistair Francis

On Apr 10 2020, Alistair Francis via Libc-alpha wrote:

> Patch 600f00b "linux: Use long time_t for wait4/getrusage" introduced
> two bugs:
>  - The usage32 struct was set if the wait4 syscall had an error.
>  - For 32-bit systems the usage struct was set even if it was specified
>    as NULL.
>
> This patch fixes the two issues.

Ok.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v4 2/2] posix: Add wait4 test case
  2020-04-10 18:55 ` [PATCH v4 2/2] posix: " Alistair Francis
@ 2020-04-14 12:08   ` Andreas Schwab
  0 siblings, 0 replies; 5+ messages in thread
From: Andreas Schwab @ 2020-04-14 12:08 UTC (permalink / raw)
  To: Alistair Francis via Libc-alpha; +Cc: Alistair Francis

On Apr 10 2020, Alistair Francis via Libc-alpha wrote:

> Add a wait4 test case based on the waitid test case.

Ok.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2020-04-14 12:08 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-04-10 18:55 [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison Alistair Francis
2020-04-10 18:55 ` [PATCH v3 2/2] linux: Add wait4 test case Alistair Francis
2020-04-10 18:55 ` [PATCH v4 2/2] posix: " Alistair Francis
2020-04-14 12:08   ` Andreas Schwab
2020-04-14 12:07 ` [PATCH v4 1/2] linux: wait4: Fix incorrect return value comparison Andreas Schwab

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).