Re: kernel provided system call is an API?

["Steve Graegert" <graegerts@gmail.com>]

On 10/29/07, Hendrik Visage <hvjunk@gmail.com> wrote:
> On 10/27/07, Glynn Clements <glynn@gclements.plus.com> wrote:
> >
> > Shriramana Sharma wrote:
> >
> > > >> 1. Can the system calls provided by the kernel be considered an API?
> > > >
> > > > System calls aren't strictly an *application* programming interface.
> > > > Applications are supposed to use the libc functions.
> > >
> > > OK so is it at least a "programming interface"?
> >
> > Yes, although it's intended to be used solely by libc.
>
> I think the "more correct"(TM) explanation would be that:
>
>  Each hardware architecture, and OS, will have different methods of
> invoking the actual system call. So from memory Linux and FreeBSD uses
> different methods to initiate a syscall.

And so do Windows, VMS and HP-UX, and many others.

> Althought the man read(2) "api" from the libc "user"'s perspective are all
> the same (specified by POSIX etc.) the nitty gritty of getting into the
> kernel differs.

Right, but we are talking about the interfaces to the _Linux_ kernel
and already illustrated in earlier threads what system calls are and
how they are invoked by making use of system call wrappers.

Anyhow, Glynn put it straight and did so in a concise manner.

	\Steve

--

Steve Grägert
DigitalEther.de
-
To unsubscribe from this list: send the line "unsubscribe linux-c-programming" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html