[uClibc]Re: vfork test fail in ppc target

Colin.Helliwell at Zarlink.Com Colin.Helliwell at Zarlink.Com
Wed Sep 4 12:51:16 UTC 2002


The command being run by the system() call does execute ok, then the whole
application (incorrectly) terminates, whereas when built against glibc it
operates as it should.
Loosely paraphrasing, the app does something like this:

while(1)
{
    if (the right thing happens)
    {
        system("echo got it");
    }
}

As with the vfork test, changing the command called to a suitable "ps"
shows the child process having a PPID of 1 (i.e. the init process!):
  PID  PPID COMMAND
23388     1 ps -ao pid,ppid,args

instead of the PID value of the parent application as with glibc:
  PID  PPID COMMAND
23389 22555 ./appln
23390 23389 ps -ao pid,ppid,args


Colin



                                                                                                                    
                    Miles Bader                                                                                     
                    <miles at gnu.org       To:     Colin.Helliwell at Zarlink.Com                                        
                    >                    cc:     uClibc at uclibc.org                                                  
                                         Subject:     Re: vfork test fail in ppc target                             
                    04-Sep-2002                                                                                     
                    01:28 PM                                                                                        
                                                                                                                    
                                                                                                                    




On Wed, Sep 04, 2002 at 11:17:23AM +0100, Colin.Helliwell at Zarlink.Com
wrote:
> > vfork and fork are _different_, and applications have to be aware of
the
> > differences.  Do you have an application [not a test program] that's
> > using vfork, and failing?
>
> Yes I do, in effect - I have an application which uses system(), which is
> implemented by uClibc via a vfork (since __NR_vfork *is* #define'd).

So what does it do wrong?

-Miles
--
P.S.  All information contained in the above letter is false,
      for reasons of military security.







More information about the uClibc mailing list