On Sun, 16 Jan 2000, Guest section DW wrote:
> >
> > The usefulness in distinquishing the two is that you know whether to look
> > at the libs or the kernel if a problem exists.
>
> No. There is no such distinction.
>
> I guess that your question is inspired by this getpagesize() business,
> so take that as example.
>
> First of all, you can test using strace whether something leads to a syscall
> or not. In the case of getpagesize() that depends on the architecture, and
> on the libc version, and on whether the program was compiled statically.
> (Usually [g]libc thinks it knows the answer itself, and hence does not
> use a system call, except possibly in the case of a Sparc under
> glibc, where the system call getpagesize() may be used. If the program
> was a dynamic ELF program then pagesize is derived by the ELF loader.
> On some systems sysconf(_SC_PAGESIZE) is used.)
>
> You see - system call? Or not? Depends on a dozen other things.
> There is no clear-cut distinction.
> It so happens that the man page is in man2. There is no better place.
I won't lie and say it makes sense to me, why it's implimented as a system
call on one architecture, library function on another, etc, and why there is no
distinction.
From my vantage point it would just seem to make troubleshooting a real
bitch.
Kernel is maintained by one group of people, libc by another. Now it is
less of a mystery to me why there are so many conflicts and things out of sync.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Jan 23 2000 - 21:00:12 EST