linux-kernel - Re: [patch 00/13] vfs: add helpers to check r/o bind mounts

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]

Message-Id: <E1Jp5AP-0004Zd-ID@pomaz-ex.szeredi.hu>
Date:	Thu, 24 Apr 2008 19:23:29 +0200
From:	Miklos Szeredi <miklos@...redi.hu>
To:	ezk@...sunysb.edu
CC:	miklos@...redi.hu, ezk@...sunysb.edu, viro@...iv.linux.org.uk,
	akpm@...ux-foundation.org, torvalds@...ux-foundation.org,
	dave@...ux.vnet.ibm.com, mhalcrow@...ibm.com,
	linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [patch 00/13] vfs: add helpers to check r/o bind mounts

> > > Sure, unionfs already does some of it now (and some of it can be recoded
> > > more cleanly), but I'd like to get rid of much of it if I could -- these new
> > > path_* helpers make me have to maintain vfsmounts even moreso than before.
> > 
> > With the words of Al: "that is hogwash".  Stacks either maintain
> > vfsmounts or they don't.  There's no middle ground.
> 
> So, if I wanted to not maintain vfsmounts at all in unionfs, how can I use
> the proposed new path_* helpers which require vfsmounts?  Will there be some
> other helpers available to perform lower-filesystem operations (e.g., mkdir,
> create, unlink, etc.) which won't require passing vfsmounts?
> 
> The "or they don't" is not much of an option when I'm forced to use an API
> that requires a vfsmount...

Yes.  The vfsmount is already needed for open() and that pretty much
determines the fate of all stacking filesystems, except the ones which
don't want to do file I/O, but that is rather hard to imagine.

And anyway, I don't see the issue here.  If the stack wants to work on
a single filesystem directly, just do a bind mount of the original to
a kernel private mount and use that.  I'm not sure this can be done
with the current API, but it doesn't sound difficult to implement at
all.

Miklos
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/