Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Song Liu <songliubraving@fb.com>
To:     Daniel Xu <dxu@dxuuu.xyz>
CC:     bpf <bpf@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "ast@kernel.org" <ast@kernel.org>,
        "daniel@iogearbox.net" <daniel@iogearbox.net>,
        Kernel Team <Kernel-team@fb.com>
Subject: Re: [PATCH bpf v2 1/2] lib/strncpy_from_user.c: Don't overcopy bytes
 after NUL terminator
Thread-Topic: [PATCH bpf v2 1/2] lib/strncpy_from_user.c: Don't overcopy bytes
 after NUL terminator
Thread-Index: AQHWsxsGDNyincyzREiD40J5/8I0oqm52XIA
Date:   Thu, 5 Nov 2020 18:16:51 +0000
Message-ID: <CE6BCF1F-2112-40DC-87C8-91FA2D6C86FC@fb.com>
References: <cover.1604542786.git.dxu@dxuuu.xyz>
 <487a07aa911b4e822a0b931f7b33a4f67fedb6bd.1604542786.git.dxu@dxuuu.xyz>
In-Reply-To: <487a07aa911b4e822a0b931f7b33a4f67fedb6bd.1604542786.git.dxu@dxuuu.xyz>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="us-ascii"
Content-ID: <FC80BB85421AD5409BBAFAE3FC1E21CE@namprd15.prod.outlook.com>
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-MS-Exchange-CrossTenant-Network-Message-Id: 00d7564b-5e81-4bdb-bd7a-08d881b6f7a5
X-MS-Exchange-CrossTenant-originalarrivaltime: 05 Nov 2020 18:16:51.0591
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Hosted
X-MS-Exchange-CrossTenant-id: 8ae927fe-1255-47a7-a2af-5f3a069daaa2
X-MS-Exchange-CrossTenant-mailboxtype: HOSTED
X-MS-Exchange-CrossTenant-userprincipalname: npZD5Zhct+WPvqjMJ/omJDXvA/zOltTmyJ19kOA3ZMIlolVgdgC4OLSiACEUH1qgyDkWB8cXbSRydqCEedWd6Q==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BYAPR15MB2568
X-OriginatorOrg: fb.com
X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:6.0.312,18.0.737
 definitions=2020-11-05_11:2020-11-05,2020-11-05 signatures=0
X-Proofpoint-Spam-Details: rule=fb_default_notspam policy=fb_default score=0 clxscore=1011 bulkscore=0
 impostorscore=0 malwarescore=0 adultscore=0 suspectscore=0 mlxscore=0
 mlxlogscore=936 priorityscore=1501 phishscore=0 spamscore=0
 lowpriorityscore=0 classifier=spam adjust=0 reason=mlx scancount=1
 engine=8.12.0-2009150000 definitions=main-2011050118
X-FB-Internal: deliver
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org


> On Nov 4, 2020, at 6:25 PM, Daniel Xu <dxu@dxuuu.xyz> wrote:
>=20
> do_strncpy_from_user() may copy some extra bytes after the NUL

We have multiple use of "NUL" here, should be "NULL"?

> terminator into the destination buffer. This usually does not matter for
> normal string operations. However, when BPF programs key BPF maps with
> strings, this matters a lot.
>=20
> A BPF program may read strings from user memory by calling the
> bpf_probe_read_user_str() helper which eventually calls
> do_strncpy_from_user(). The program can then key a map with the
> resulting string. BPF map keys are fixed-width and string-agnostic,
> meaning that map keys are treated as a set of bytes.
>=20
> The issue is when do_strncpy_from_user() overcopies bytes after the NUL
> terminator, it can result in seemingly identical strings occupying
> multiple slots in a BPF map. This behavior is subtle and totally
> unexpected by the user.
>=20
> This commit uses the proper word-at-a-time APIs to avoid overcopying.
>=20
> Fixes: 6ae08ae3dea2 ("bpf: Add probe_read_{user, kernel} and probe_read_{=
user, kernel}_str helpers")
> Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
> ---
> lib/strncpy_from_user.c | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)
>=20
> diff --git a/lib/strncpy_from_user.c b/lib/strncpy_from_user.c
> index e6d5fcc2cdf3..d084189eb05c 100644
> --- a/lib/strncpy_from_user.c
> +++ b/lib/strncpy_from_user.c
> @@ -35,17 +35,22 @@ static inline long do_strncpy_from_user(char *dst, co=
nst char __user *src,
> 		goto byte_at_a_time;
>=20
> 	while (max >=3D sizeof(unsigned long)) {
> -		unsigned long c, data;
> +		unsigned long c, data, mask, *out;
>=20
> 		/* Fall back to byte-at-a-time if we get a page fault */
> 		unsafe_get_user(c, (unsigned long __user *)(src+res), byte_at_a_time);
>=20
> -		*(unsigned long *)(dst+res) =3D c;
> 		if (has_zero(c, &data, &constants)) {
> 			data =3D prep_zero_mask(c, data, &constants);
> 			data =3D create_zero_mask(data);
> +			mask =3D zero_bytemask(data);
> +			out =3D (unsigned long *)(dst+res);
> +			*out =3D (*out & ~mask) | (c & mask);
> 			return res + find_zero(data);
> +		} else  {

This else clause is not needed, as we return in the if clause.=20

> +			*(unsigned long *)(dst+res) =3D c;
> 		}
> +
> 		res +=3D sizeof(unsigned long);
> 		max -=3D sizeof(unsigned long);
> 	}
> --=20
> 2.28.0
>=20