linux-kernel - Re: [PATCH 12/15] VFS: Add owner-filesystem positive/negative dentry checks

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:	Fri, 27 Mar 2015 16:27:28 +0100
From:	Miklos Szeredi <miklos@...redi.hu>
To:	David Howells <dhowells@...hat.com>
Cc:	viro@....linux.org.uk,
	"linux-unionfs@...r.kernel.org" <linux-unionfs@...r.kernel.org>,
	Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 12/15] VFS: Add owner-filesystem positive/negative dentry checks

On Fri, Mar 27, 2015 at 3:42 PM, David Howells <dhowells@...hat.com> wrote:
> Miklos Szeredi <miklos@...redi.hu> wrote:
>
>> I think this is confusing as hell, there needs to be more consistency
>> in the naming.  E.g. d_backing_is_positive() vs. d_is_positive().   I
>> know it's the other way round now, but only with a few users.
>
> Yeah.  The problem is that all of:
>
>         __d_entry_type()
>         d_is_miss()
>         d_is_whiteout()
>         d_can_lookup()
>         d_is_autodir()
>         d_is_dir()
>         d_is_symlink()
>         d_is_reg()
>         d_is_special()
>         d_is_file()
>         d_is_negative()
>         d_is_positive()
>
> refer to the 'backing' inode (if there is one) in the case that you have a
> unionmount and the top dentry's ->d_inode is NULL.  (Well, technically, that
> doesn't happen in the case of directories)

Looks to me we actually need two variants of all of the above, since
most filesystems never want to refer to the backing inode.

>
> Of course, if we decide we aren't going to do unionmount, certain things
> become simpler.
>
>> Also a separate include file might help, that needs explicit including to
>> get the "backing" variants
>
> I would like to see a 'for fs implementer' header and a 'for fs user' header
> but Al didn't like that last time I suggested it.
>
> However, it doesn't help with the naming since there are situations where you
> need *both* - eg. overlayfs.

Not sure what you mean, the naming *must* be different even if we have
two headers and overlayfs can just include both.

Thanks,
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/