Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <d955d8b5-ca2e-c040-9415-772fa5a71bc7@schaufler-ca.com>
Date:   Thu, 1 Sep 2022 09:45:51 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Thunderbird/91.13.0
Subject: Re: [RFC PATCH 1/2] fs/xattr: add *at family syscalls
Content-Language: en-US
To:     Al Viro <viro@zeniv.linux.org.uk>,
        =?UTF-8?Q?Christian_G=c3=b6ttsche?= <cgzones@googlemail.com>
Cc:     selinux@vger.kernel.org, linux-fsdevel@vger.kernel.org,
        linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org,
        Luis Chamberlain <mcgrof@kernel.org>,
        LSM List <linux-security-module@vger.kernel.org>
References: <20220830152858.14866-1-cgzones@googlemail.com>
 <20220830152858.14866-2-cgzones@googlemail.com> <Yw/eEufm/QpKg5Pq@ZenIV>
From:   Casey Schaufler <casey@schaufler-ca.com>
In-Reply-To: <Yw/eEufm/QpKg5Pq@ZenIV>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 8/31/2022 3:17 PM, Al Viro wrote:
> [linux-arch Cc'd for ABI-related stuff]

The LSM list <linux-security-module@vger.kernel.org> should be on
this thread as SELinux isn't the only security module that uses xattrs
extensively.

>
> On Tue, Aug 30, 2022 at 05:28:39PM +0200, Christian Göttsche wrote:
>> Add the four syscalls setxattrat(), getxattrat(), listxattrat() and
>> removexattrat() to enable extended attribute operations via file
>> descriptors.  This can be used from userspace to avoid race conditions,
>> especially on security related extended attributes, like SELinux labels
>> ("security.selinux") via setfiles(8).
>>
>> Use the do_{name}at() pattern from fs/open.c.
>> Use a single flag parameter for extended attribute flags (currently
>> XATTR_CREATE and XATTR_REPLACE) and *at() flags to not exceed six
>> syscall arguments in setxattrat().
> 	I've no problems with the patchset aside of the flags part;
> however, note that XATTR_CREATE and XATTR_REPLACE are actually exposed
> to the network - the values are passed to nfsd by clients.
> See nfsd4_decode_setxattr() and
>         BUILD_BUG_ON(XATTR_CREATE != SETXATTR4_CREATE);
> 	BUILD_BUG_ON(XATTR_REPLACE != SETXATTR4_REPLACE);
> in encode_setxattr() on the client side.
>
> 	Makes me really nervous about constraints like that.  Sure,
> AT_... flags you are using are in the second octet and these are in
> the lowest one, but...