Date: Wed, 30 Jul 2003 21:52:48 -0700 From: Bill Fenner <fenner@research.att.com> To: kris@obsecurity.org Cc: standards@freebsd.org Subject: Re: struct timeval Message-ID: <200307310452.h6V4qn402047@windsor.research.att.com> References: <20030615053705.GA15421@rot13.obsecurity.org> <20030729000206.GA92727@rot13.obsecurity.org> <20030728234338.B94307@espresso.q9media.com> <20030729042506.GA61736@rot13.obsecurity.org>
next in thread | previous in thread | raw e-mail | index | archive | help
I think we can do this with old-binary-on-new-kernel compatability,
by putting in explicit padding that can be zeroed by the kernel.
new-binary-on-old-kernel should also work, since the explicit padding
in the struct can put the time_t in the right place.
My strawman is something like:
struct timeval {
#ifdef __sparc64__
uint32_t __pad;
#endif
time_t tv_sec;
#ifdef __alpha__
uint32_t __pad;
#endif
suseconds_t tv_usec;
}
and anything that returns a timeval has to
#if defined(__sparc64__) || defined(__alpha__)
tv->__pad = 0;
#endif
POSIX 1003.1-2001 says:
sys/time.h - time types
#include <sys/time.h>
The <sys/time.h> header shall define the timeval structure that includes
at least the following members:
time_t tv_sec Seconds.
suseconds_t tv_usec Microseconds.
so the optional padding is OK and the type of tv_sec is correct.
Bill
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200307310452.h6V4qn402047>