linux-kernel - Re: [PATCH] fs/ceph/super: add mount options "snapdir{mode,uid,gid}"

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] [day] [month] [year] [list]

Date:   Tue, 25 Oct 2022 11:57:59 +0200
From:   Max Kellermann <max.kellermann@...os.com>
To:     Xiubo Li <xiubli@...hat.com>
Cc:     Jeff Layton <jlayton@...nel.org>, idryomov@...il.com,
        ceph-devel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] fs/ceph/super: add mount options "snapdir{mode,uid,gid}"

On Tue, Oct 25, 2022 at 11:10 AM Xiubo Li <xiubli@...hat.com> wrote:
> $ sudo ./bin/mount.ceph privileged@.../ /mnt/privileged/mountpoint
>
> $ sudo ./bin/mount.ceph global@.../ /mnt/global/mountpoint

So you have two different mount points where different client
permissions are used. There are various problems with that
architecture:

- it complicates administration, because now every mount has to be done twice
- it complicates applications accessing ceph (and their
configuration), because there are now 2 mount points
- it increases resource usage for having twice as many ceph connections
- it interferes with fscache, doubling fscache's local disk usage,
reducing fscache's efficiency
- ownership of the snapdir is still the same as the parent directory,
and I can't have non-superuser processes to manage snapshots; all
processes mananging snapshots need to have write permission on the
parent directory
- this is still all-or-nothing; I can't forbid users to list (+r) or
access (+x) snapshots

All those problems don't exist with my patch.