[Valgrind-users] [ BUG ] pthread_once

[Steve G <lin...@ya...>]

Hello,

I ran across a problem auditing a daemon. Appearantly, it
calls pthread_once() more than once and valgrind terminates
the run. In coregrind/vg_libpthread.c around line 1500 is
this code:

   res = __pthread_mutex_lock(&once_masterlock);
                                                           
                    
   if (res != 0) {
      barf("pthread_once: Looks like your program's "
           "init routine calls back to pthread_once() ?!");
   }
                                                           
                    
The man page for pthread_once says: 

Subsequent calls to pthread_once  with  the  same 
once_control argument do nothing.

So, I changed "barf()" to "return 0" and everything works
OK now. It seems that valgrind ought to store the
init_routine's address away and just return if they've
already been called/stored. FWIW, the daemon I'm testing is
/usr/sbin/named.

Best Regards,
Steve Grubb

__________________________________
Do you Yahoo!?
Yahoo! Calendar - Free online calendar with sync to Outlook(TM).
http://calendar.yahoo.com