[PATCH] More stackgap removal

John Baldwin jhb at FreeBSD.org
Wed Mar 9 22:42:57 GMT 2005


On Wednesday 09 March 2005 04:56 pm, Jung-uk Kim wrote:
> On Wednesday 09 March 2005 04:52 pm, Jung-uk Kim wrote:
> > On Wednesday 09 March 2005 10:31 am, John Baldwin wrote:
> > > On Wednesday 02 March 2005 04:43 pm, John Baldwin wrote:
> > > > I've fixed up a few more syscalls to have kern_foo versions so
> > > > that ABI wrappers for them can stop using the stackgap.  In
> > > > this patch, I've added kern_settimeofday() for ibcs2 and
> > > > freebsd32, kern_adjtime() for freebsd32, and
> > > > kern_sched_rr_get_interval() for Linux/i386 on amd64 (note that
> > > > this means that COMPAT_LINUX32 now requires
> > > > POSIX_PRIORITY_SCHEDULING). I've also added kern_readv() and
> > > > kern_writev() for OSF/1, Linux/i386 on amd64, and freebsd32.
> > > > For these functions I added a per-ABI version of copyinuio() to
> > > > build a uio structure from the iovec array in userland. Testing
> > > > and feedback of both native and compat binaries greatly
> > > > appreciated.
> > > >
> > > > http://www.FreeBSD.org/~jhb/patches/abi.patch
> > >
> > > Is anyone willing to test this?
> >
> > I tested freebsd32.  kern_settimeofday(), kern_adjtime(), and
> > kern_readv() seemed okay.  However, kern_writev() always fails with
> > EBADF.  Any idea?
>
> I am sorry to answer quickly to myself but I found a typo in the
> patch.  freebsd32_writev() calls kern_readv().  Oops. ;-)

Doh, I noticed that when I did that in the other ABIs. :-P  Thanks very much.

-- 
John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org


More information about the freebsd-amd64 mailing list