Re: [PATCH v8 3/4] fs: move S_ISGID stripping into the vfs

From: Christian Brauner <brauner@kernel.org>
To: Al Viro <viro@zeniv.linux.org.uk>
Cc: Yang Xu <xuyang2018.jy@fujitsu.com>,
	Dave Chinner <david@fromorbit.com>,
	"Darrick J. Wong" <djwong@kernel.org>,
	Matthew Wilcox <willy@infradead.org>,
	Jeff Layton <jlayton@kernel.org>,
	Miklos Szeredi <miklos@szeredi.hu>,
	Amir Goldstein <amir73il@gmail.com>,
	linux-fsdevel <linux-fsdevel@vger.kernel.org>,
	ceph-devel <ceph-devel@vger.kernel.org>
Subject: Re: [PATCH v8 3/4] fs: move S_ISGID stripping into the vfs
Date: Thu, 28 Apr 2022 10:07:08 +0200	[thread overview]
Message-ID: <20220428080708.y76yhqwczwwmdvi4@wittgenstein> (raw)
In-Reply-To: <Ymob0U33QNeJEeFs@zeniv-ca.linux.org.uk>

On Thu, Apr 28, 2022 at 04:45:05AM +0000, Al Viro wrote:
> On Wed, Apr 27, 2022 at 11:22:01AM +0200, Christian Brauner wrote:
> 
> > +static inline umode_t vfs_prepare_mode(struct user_namespace *mnt_userns,
> > +				       const struct inode *dir, umode_t mode,
> > +				       umode_t mask_perms, umode_t type)
> > +{
> > +	/*
> > +	 * S_ISGID stripping depends on the mode of the new file so make sure
> > +	 * that the caller gives us this information and splat if we miss it.
> > +	 */
> > +	WARN_ON_ONCE((mode & S_IFMT) == 0);
> 
> <blink>
> 
> First of all, what happens if you call mknod("/tmp/blah", 0, 0)?  And the only
> thing about type bits we care about is "is it a directory" - the sensitive
> stuff is in the low 12 bits...  What is that check about?

Do note that this is just an untested rough sketch to illustrate how to
move it into vfs_*() helpers.