linux-kernel - Re: [RFC PATCH] vfs: syscalls: Add create_automount() and remove

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 19 Feb 2020 15:43:24 +0100
From:   Geert Uytterhoeven <geert@...ux-m68k.org>
To:     David Howells <dhowells@...hat.com>
Cc:     Al Viro <viro@...iv.linux.org.uk>, coda@...cmu.edu,
        linux-afs@...r.kernel.org, linux-cifs@...r.kernel.org,
        "open list:NFS, SUNRPC, AND..." <linux-nfs@...r.kernel.org>,
        linux-btrfs <linux-btrfs@...r.kernel.org>,
        Linux FS Devel <linux-fsdevel@...r.kernel.org>,
        "torvalds@...ux-foundation.org" <torvalds@...ux-foundation.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [RFC PATCH] vfs: syscalls: Add create_automount() and remove_automount()

Hi David,

On Wed, Feb 19, 2020 at 3:36 PM David Howells <dhowells@...hat.com> wrote:
> Add system calls to create and remove mountpoints().  These are modelled
> after mkdir and rmdir inside the VFS.  Currently they use the same security
> hooks which probably needs fixing.
>
> The calls look like:
>
>  long create_mountpoint(int dfd, const char *path,
>                         const char *fstype, const char *source,
>                         const char *params);
>  long remove_mountpoint(int dfd, const char *path);
>
> Creation takes an fstype, source and params which the filesystem that owns
> the mountpoint gets to filter/interpret.  It is free to reject any
> combination of fstype, source and params it cannot store.  source and
> params are both optional.
>
> Removal could probably be left to rmdir(), but this gives the option of
> applying tighter security checks and also allows me to prevent rmdir from
> removing them by accident.
>
> The AFS filesystem is then altered to use these system calls to create and
> remove persistent mountpoints in an AFS volume.  create_automount() is
> something that AFS needs, but cannot be implemented with, say, symlink().
> These substitute for the lack of pioctl() on Linux, supplying the
> functionality of VIOC_AFS_CREATE_MT_PT and VIOC_AFS_DELETE_MT_PT.
>
> Also make them usable with tmpfs for testing.  I'm not sure if this is
> useful in practice, but I've made tmpfs store the three parameters and just
> pass them to mount when triggered.  Note that it doesn't look up the target
> filesystem until triggered so as not to load lots of modules until
> necessary.
>
> I suspect they're of little of use to NFS, CIFS and autofs, but probably
> Coda and maybe Btrfs can make use of them.
>
> Signed-off-by: David Howells <dhowells@...hat.com>

Thanks for your patch!

The above nicely explains what the patch does.
However, unless I'm missing something, this fails to explain the "why"
(except for the vague "[...] is something that AFS needs ...".

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds