From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-fsdevel-owner@vger.kernel.org>
Received: from smtp.gentoo.org ([140.211.166.183]:48186 "EHLO smtp.gentoo.org"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1752794AbcHQUhr (ORCPT <rfc822;linux-fsdevel@vger.kernel.org>);
	Wed, 17 Aug 2016 16:37:47 -0400
Date: Wed, 17 Aug 2016 13:37:46 -0700
From: Mike Frysinger <vapier@gentoo.org>
To: Jeff Layton <jlayton@redhat.com>
Cc: "Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>,
	libc-alpha@sourceware.org, linux-fsdevel@vger.kernel.org,
	Carlos O'Donell <carlos@redhat.com>,
	Yuriy Kolerov <Yuriy.Kolerov@synopsys.com>
Subject: Re: [glibc PATCH] fcntl: put F_OFD_* constants under #ifdef
 __USE_FILE_OFFSET64
Message-ID: <20160817203746.GF21655@vapier.lan>
References: <1471445251-2450-1-git-send-email-jlayton@redhat.com>
 <20160817184333.GC21655@vapier.lan>
 <1471461304.3196.101.camel@redhat.com>
 <b470dd43-e77d-62ff-1b5d-2535febb2eb1@gmail.com>
 <1471464343.3196.125.camel@redhat.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha256;
	protocol="application/pgp-signature"; boundary="OzxllxdKGCiKxUZM"
Content-Disposition: inline
In-Reply-To: <1471464343.3196.125.camel@redhat.com>
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>


--OzxllxdKGCiKxUZM
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On 17 Aug 2016 16:05, Jeff Layton wrote:
> The way it works now is that when you define _FILE_OFFSET_BITS=3D64 and
> call fcntl(fd, F_SETLK, fl) glibc swaps in a struct flock64 for your
> struct flock, and F_SETLK64 for the F_SETLK.

does it ?  doesn't seem like it does to me.  here's glibc's fcntl.c:
	io/fcntl.c - generic stub that sets ENOSYS
	sysdeps/unix/sysv/linux/fcntl.c - just calls syscall(fcntl)
	sysdeps/unix/sysv/linux/generic/wordsize-32/fcntl.c - just calls syscall(f=
cntl64)
	sysdeps/unix/sysv/linux/i386/fcntl.c - same as above
	<all the other 32-bit arches include the i386 file>

the kernel is where it gets interesting:
	fs/compat.c:
		COMPAT_SYSCALL_DEFINE3(fcntl):
			rejects all 64-bit commands w/EINVAL
			passes all other calls to compat_sys_fcntl64
		COMPAT_SYSCALL_DEFINE3(fcntl64):
			rewrites 32-bit flock struct to 64-bit flock struct
			passes args to sys_fcntl=20
	fs/fcntl.c:
		SYSCALL_DEFINE3(fcntl):
			passes all args to do_fcntl
		SYSCALL_DEFINE3(fcntl64):
			handles 64-bit flock commands
			passes all others commands to do_fcntl
		do_fcntl:
			handles all commands using native sized flock struct

so for a 32-bit system (e.g. i386), you must match LFS & command usage.
if LFS is turned on, then using 32-bit commands w/struct flock fails.
if LFS is turned off, then using 64-bit commands w/struct flock fails.
-mike

--OzxllxdKGCiKxUZM
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=ms3q
-----END PGP SIGNATURE-----

--OzxllxdKGCiKxUZM--