On Tue, Jul 21, 2009 at 04:05:02PM +0200, Corinna Vinschen wrote: >On Jul 21 15:58, Corinna Vinschen wrote: >> On Jul 21 07:04, Eric Blake wrote: >> > -----BEGIN PGP SIGNED MESSAGE----- >> > Hash: SHA1 >> > >> > It seems like it would be pretty easy to add several new close-on-exec >> > features required by POSIX 2008: >> > >> > open(name, O_RDONLY | O_CLOEXEC); >> > // skips the need to use fcntl >> > >> > fcntl(fd, F_DUPFD_CLOEXEC); >> > // like F_DUPFD, but closes window without having to use >> > // fcntl(fd, F_SETFD, fcntl(fd, F_GETFD)) >> > >> > and those provided by glibc as extensions: >> > >> > fopen(name, "re"); >> > // like fopen "r", but with cloexec set >> > >> > The argument is that in multi-threaded apps, if one thread opens a file >> > but has not yet turned on the close-on-exec bit while another thread does >> > a fork-and-exec, then you leaked the fd into the child; and these new >> > flags close the window. >> > >> > Should I go ahead and prepare a patch for the newlib side? >> >> It's not that easy to implement, especially not the fcntl which requires >> to redefine the dup2 methods throughout to allow atomic operation on the >> target OS handles. open (O_CLOEXEC) is much easier. Anyway, can we >> wait until after Cygwin 1.7.1? > >...which is to say, I added it to my TODO list.
I agree about waiting, but I think this is one for my TODO list. The close_on_exec code is all mine, AFAIK. I reworked it all back in 1998 or so, so it is pretty fresh in my mind. Coincidentally, enough, I've been thinking about reworking it lately after having to deal with it for firos. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
