open() return value

Bernhard Reutner-Fischer rep.dot.nop at gmail.com
Fri Jul 13 20:44:32 UTC 2012


On Jul 13, 2012 3:08 PM, "Qais Yousef" <Qais.Yousef at imgtec.com> wrote:
>
> Thanks. I checked the syscalls.h and it is setup correctly. I did extra
testing and debugging and found out that the failure happens only in a
particular case in my application. A test app doesn't fail, and moving the
call to open() to a different location in my app behaves correctly too. It
seems to be a threading related issue.

Can you reproduce which sequence of calls trigger this behaviour? Can you
reproduce it with current master? What arch do you use and which version of
uClibc?
Thanks,
>
> One thing. I was expecting a call to open() to use open64(), but isn't
(breakpoint triggers normal open()). Open64() has some thread related
checking/tidying while open() hasn't. I wonder if this is causing the
problem? How can I make sure that open64() is used?
>
> Thanks,
> Qais
>
> > -----Original Message-----
> > From: liaoxinglong at icubecorp.com [mailto:liaoxinglong at icubecorp.com]
> > Sent: 12 July 2012 10:29
> > To: Qais Yousef
> > Subject: Re: open() return value
> >
> > open() as a system call, it will call _syscall3, and _syscall3
implement as a
> > macro in libc/sysdeps/linux/common/bits/syscalls-common.h. at the end,
in
> > macro INLINE_SYSCALL_NCS, you can see __res = -1L; but this file may be
not
> > really use, you can see syscalls.h depend on your arch. as example, you
can
> > find libc/sysdeps/linux/mips/bits/syscalls.h in MIPS arch.
> > hope this can help you.
> >
> >
> > xinglong
> >
> >
> > >  -------Original Message-------
> > >  From: Qais Yousef <Qais.Yousef at imgtec.com>
> > >  To: uClibc at uclibc.org <uClibc at uclibc.org>
> > >  Subject: open() return value
> > >  Sent: Jul 12 '12 17:06
> > >
> > >  Hi,
> > >
> > >  I'm not well versed with what the standard says, but looking at the
man
> > page of open() function it states that it returns -1 and sets errno,
but uClibc
> > implementation returns errno directly.
> > >
> > >  Not a problem for me to be honest, but I just thought I'd point it
> > > out :)
> > >
> > >  Thanks,
> > >  Qais
> > >
> > >  _______________________________________________
> > >  uClibc mailing list
> > >  uClibc at uclibc.org
> > >  http://lists.busybox.net/mailman/listinfo/uclibc
> > >
> _______________________________________________
> uClibc mailing list
> uClibc at uclibc.org
> http://lists.busybox.net/mailman/listinfo/uclibc


More information about the uClibc mailing list