From mboxrd@z Thu Jan  1 00:00:00 1970
From: Eric Dumazet <eric.dumazet@gmail.com>
Subject: Re: [PATCH v2] fs: select: fix information leak to userspace
Date: Tue, 23 Nov 2010 01:20:48 +0100
Message-ID: <1290471649.2704.24.camel@edumazet-laptop>
References: <1289421483-23907-1-git-send-email-segooon@gmail.com>
	 <20101112120834.33062900.akpm@linux-foundation.org>
	 <8D90F8B2-EA29-4EB9-9807-294CE0D5523B@dilger.ca>
	 <20101114092533.GB5323@albatros>
	 <20101114180643.593d19ac.akpm@linux-foundation.org>
	 <1289848341.2607.125.camel@edumazet-laptop> <4CE268C8.5010203@panasas.com>
	 <20101122155043.fbbb74f4.akpm@linux-foundation.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: Boaz Harrosh <bharrosh@panasas.com>,
	Vasiliy Kulikov <segoon@openwall.com>,
	Andreas Dilger <adilger.kernel@dilger.ca>,
	kernel-janitors@vger.kernel.org,
	Alexander Viro <viro@zeniv.linux.org.uk>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	Jakub Jelinek <jakub@redhat.com>
To: Andrew Morton <akpm@linux-foundation.org>
Return-path: <linux-kernel-owner@vger.kernel.org>
In-Reply-To: <20101122155043.fbbb74f4.akpm@linux-foundation.org>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org

Le lundi 22 novembre 2010 =C3=A0 15:50 -0800, Andrew Morton a =C3=A9cri=
t :

> Well.  We certainly assume in many places that
>=20
> 	struct foo {
> 		int a;
> 		int b;
> 	} f =3D {
> 		.a =3D 1,
> 	};
>=20
> will initialise b to zero.  But I doubt if much code at all assumes
> that this initialisation patterm will reliably zero out *holes* in th=
e
> struct.
>=20

We did such assertions in the past, we were wrong.

Check commit 1c40be12f7d8ca1d387510d39787b12e512a7ce8 for an example
(net sched: fix some kernel memory leaks)

I guess we must make a full audit of all C99 initializers or structures
copied to userspace, giving a name to hidden holes, to force gcc to ini=
t
them to 0.

# cat try.c
struct s {
	char c;
	long l;
	};

void bar(void *v)
{
	unsigned long *p =3D v;

	printf("%lx %lx\n", p[0], p[1]);
}

int main()
{
	struct s s1 =3D {
		.c =3D 1,
		.l =3D 2,
	};

	bar(&s1);
	return 0;
}

# gcc -O2 -o try try.c
# ./try
8049401 2

Strangely, if we remove ".l =3D 2," line, gcc emits code to clear al th=
e
fields

main:
	pushl	%ebp
	movl	%esp, %ebp
	andl	$-16, %esp
	subl	$32, %esp
	leal	24(%esp), %eax
	movl	$0, 24(%esp)
	movl	%eax, (%esp)
	movl	$0, 28(%esp)
	movb	$1, 24(%esp)
	call	bar
	xorl	%eax, %eax
	leave
	ret