This is the mail archive of the libc-hacker@sourceware.org mailing list for the glibc project.

Note that libc-hacker is a closed list. You may look at the archives of this list, but subscription and posting are not open.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[PATCH] Fix tst-pselect failure handling


Hi!

When tst-pselect.c fails (quite often when running on kernel that
doesn't have pselect6 syscall, because pselect is then not atomic,
but even 2.6.18 x86_64 kernel it fails rather quickly with pselect6
syscall when running while misc/tst-pselect; do :; done), test-skeleton.c
only kills its immediate child, while the other child sticks around
forever.

The following patch is one possibility to avoid this:

2006-12-14  Jakub Jelinek  <jakub@redhat.com>

	* misc/tst-pselect.c (do_test): Break the child loop if
	parent was killed by timeout handler.

--- libc/misc/tst-pselect.c.jj	2006-09-18 09:11:59.000000000 +0200
+++ libc/misc/tst-pselect.c	2006-12-14 16:35:33.000000000 +0100
@@ -61,12 +61,15 @@ do_test (void)
       FD_SET (fds[0][0], &rfds);
 
       int e;
+      /* Break the loop if our parent is killed by test-skeleton.c's
+	 timeout handler.  */
+      pid_t ppid = getppid ();
       do
 	{
 	  errno = 0;
 	  e = pselect (fds[0][0] + 1, &rfds, NULL, NULL, &to, &ss);
 	}
-      while (e == 0);
+      while (e == 0 && getppid () == ppid);
 
       if (e != -1)
 	{

	Jakub


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]