From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-fsdevel-owner@vger.kernel.org>
Received: from mail-oi0-f67.google.com ([209.85.218.67]:36754 "EHLO
        mail-oi0-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1754515AbdCILNb (ORCPT
        <rfc822;linux-fsdevel@vger.kernel.org>);
        Thu, 9 Mar 2017 06:13:31 -0500
Received: by mail-oi0-f67.google.com with SMTP id a144so4339515oib.3
        for <linux-fsdevel@vger.kernel.org>; Thu, 09 Mar 2017 03:13:30 -0800 (PST)
MIME-Version: 1.0
In-Reply-To: <20170219032743.GH29622@ZenIV.linux.org.uk>
References: <1484588765-9397-1-git-send-email-amir73il@gmail.com>
 <1484588765-9397-2-git-send-email-amir73il@gmail.com> <CAJfpegsDzNot=z8oP7DecuXtNkYPYOejDWWXs6KjD6=Le=Wohg@mail.gmail.com>
 <20170219032743.GH29622@ZenIV.linux.org.uk>
From: Miklos Szeredi <miklos@szeredi.hu>
Date: Thu, 9 Mar 2017 12:13:29 +0100
Message-ID: <CAJfpegsLHq8qhv+nBrgF3f0tM_LDLoOk1vTxpsdnwnc78hcppA@mail.gmail.com>
Subject: Re: [PATCH v3 1/6] vfs: create vfs helper vfs_tmpfile()
To: Al Viro <viro@zeniv.linux.org.uk>
Cc: Amir Goldstein <amir73il@gmail.com>,
        "linux-unionfs@vger.kernel.org" <linux-unionfs@vger.kernel.org>,
        linux-fsdevel@vger.kernel.org,
        "Eric W. Biederman" <ebiederm@xmission.com>
Content-Type: text/plain; charset=UTF-8
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Sun, Feb 19, 2017 at 4:27 AM, Al Viro <viro@zeniv.linux.org.uk> wrote:
> On Mon, Jan 16, 2017 at 08:47:32PM +0100, Miklos Szeredi wrote:
>
>> > +                          umode_t mode, int open_flag)
>> > +{
>> > +       static const struct qstr name = QSTR_INIT("/", 1);
>> > +       struct dentry *child = NULL;
>> > +       struct inode *inode;
>> > +       int error;
>> > +
>> > +       /* we want directory to be writable */
>> > +       error = inode_permission(dir, MAY_WRITE | MAY_EXEC);
>>
>> This is not in the scope of this patch, but shoudln't we be using
>> may_create() here?   Or at least a variant without the audit thing...
>>
>> Al?
>
> may_create() expects directory + child dentry; here we have only parent.
> IS_DEADDIR is rather pointless here - directory is not locked, for
> starters, so rmdir might happen right under you.  Or right after you've
> returned from your function, for that matter.  userns checks...
> FWIW, no such checks are done in ->atomic_open() paths, so I'm not sure
> how much are those worth...

Eric would know since he added those checks.

Thanks,
Miklos