linux-kernel - Re: [PATCH v2] vfs: Tighten up linkat(..., AT_EMPTY

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:	Thu, 22 Aug 2013 20:53:17 +0200
From:	Willy Tarreau <w@....eu>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Andy Lutomirski <luto@...capital.net>,
	"security@...nel.org" <security@...nel.org>,
	Ingo Molnar <mingo@...nel.org>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Oleg Nesterov <oleg@...hat.com>,
	Al Viro <viro@...iv.linux.org.uk>,
	Linux FS Devel <linux-fsdevel@...r.kernel.org>,
	Brad Spengler <spender@...ecurity.net>
Subject: Re: [PATCH v2] vfs: Tighten up linkat(..., AT_EMPTY_PATH)

On Thu, Aug 22, 2013 at 11:48:10AM -0700, Linus Torvalds wrote:
> On Wed, Aug 21, 2013 at 12:14 PM, Andy Lutomirski <luto@...capital.net> wrote:
> >
> > So let's be careful for now: only allow linkat(..., AT_EMPTY_PATH)
> > if the target is I_LINKABLE.
> 
> So I really detest this just because it's such a special case. Now
> this is only useful for that one special case, and the thing very
> fundamentally checks that one special case in a place that is
> impossible to check for the /proc case, so the proc case remains
> totally separate.
> 
> Which just bothers me.
> 
> I think we could easily at least allow the "file->f_creds ==
> current->creds" case (and yes, I literally mean comparing the pointers
> - not only is it cheaper, but it literally means "nothing odd has
> happened in between opening and the lookup").
> 
> And I'm wondering if we shouldn't actually do that at "path_init"
> time. Right now the code says:
> 
>                 /* Caller must check execute permissions on the
> starting path component */
>                 struct fd f = fdget_raw(dfd);
> 
> and then uses the struct file mindlessly.
> 
> I'm wondering if we should just do some validation in that place, and say:
> 
>  - for directories, we require exec permissions here
>  - for everything else, we require that f->f_cred == current->cred check.
> 
> I dunno. But that I_LINKABLE thing just bothers me. It screams "I'm
> hacky" to me.

I agreed, simply because the condition here is different from the one in /proc.

I have read some code last evening to try to understand how /proc/pid/fd
entries were granted access to various processes, because I would love to
see the same condition being used in both places. Unfortunately, it's beyond
my skills, and I stopped after my random attempts gave me some panics.

Willy

--
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/