Subject: Re: Large inodes for ffs
To: Chris G. Demetriou <cgd@netbsd.org>
From: Bill Studenmund <wrstuden@nas.nasa.gov>
List: tech-kern
Date: 03/24/1999 14:15:53
On 24 Mar 1999, Chris G. Demetriou wrote:

> Jason Thorpe <thorpej@nas.nasa.gov> writes:
> 
> > Unfortunately, fcntl()'s interface is totally, completely, and utterly
> > broken.
> > 
> > int	fcntl(int fd, int cmd, int arg);
> > 
> > That "int arg" is the killer; can't pass a pointer to data.
> 
> uh, that doesn't match either the prototype or syscall declaration on
> NetBSD, though it does seem to match the NetBSD manual page.
> 
> further, if your argument were correct and the manual page actually
> matched reality, existing fcntl() in NetBSD couldn't work properly,
> since the locking-related controls pass in a pointer.  8-)
> 
> 
> Looks to me like our fcntl() manual page is broken in several ways,
> and should be updated to look more like the solaris manual page (which
> provides the correct understanding of the nature of 'arg').  8-)

Why is the declaration (in syscalls.master) "int sys_fcntl(int fd, int
cmd, ... void *arg);" - the dots make it look like it's varyarg'd when it
isn't.

More in a reply to your earlier message. :-)

Take care,

Bill