[PATCH] _syscall6 for 2.6

Steve Munroe sjmunroe at us.ibm.com
Wed Nov 26 14:11:22 EST 2003

GLIBC does have its own syscall macros and does not depend on the unistd.h
versions. I believe that GLIBCs syscalls are correct for the management of
volatile management.

Unfortunately the Kernel's syscall macros are used by the LTP in many of
their kernel level tests. We can't change this without annoying a lot of

Steven J. Munroe
Power Linux Toolchain Architect
IBM Corporation, Linux Technology Center

[ linas at austin.ibm.com writes: ]
> On Tue, Nov 25, 2003 at 02:26:18PM +0100, Franz Sirl wrote:
> >
> > why do we need this in a 2.6 kernel? Can't we call everything directly
> > now in-kernel? And using this _syscallN stuff in userspace is
> > deprecated AFAIK and if there was some consensus across architectures,
> > we could remove them completely.
> Out of curiosity, what is this replaced by? Is the syscall ABI
> sufficiently spec'ed out so that glibc can safely "guess" the right way
> to make a syscall? (Since I thought glibc used _syscallN, or does it
> have its own macros?)

** Sent via the linuxppc64-dev mail list. See http://lists.linuxppc.org/

More information about the Linuxppc64-dev mailing list