From: David Gibson <david@gibson.dropbear.id.au>
To: Stefano Brivio <sbrivio@redhat.com>
Cc: "Richard W.M. Jones" <rjones@redhat.com>,
passt-dev@passt.top, Minxi Hou <mhou@redhat.com>
Subject: Re: [PATCH 7/8] conf, passt, tap: Open socket and PID files before switching UID/GID
Date: Fri, 21 Jun 2024 11:02:38 +1000 [thread overview]
Message-ID: <ZnTRLtW8uQ3FefgT@zatzit> (raw)
In-Reply-To: <20240620162212.09ccc2f9@elisabeth>
[-- Attachment #1: Type: text/plain, Size: 3030 bytes --]
On Thu, Jun 20, 2024 at 04:22:12PM +0200, Stefano Brivio wrote:
> On Thu, 20 Jun 2024 13:47:31 +0100
> "Richard W.M. Jones" <rjones@redhat.com> wrote:
>
> > On Thu, Jun 20, 2024 at 02:12:53PM +0200, Stefano Brivio wrote:
> > > On Thu, 20 Jun 2024 12:30:54 +0100
> > > "Richard W.M. Jones" <rjones@redhat.com> wrote:
> > >
> > > > On Wed, May 29, 2024 at 12:35:24PM +1000, David Gibson wrote:
> > > > > On Wed, May 22, 2024 at 10:59:10PM +0200, Stefano Brivio wrote:
> > > > > > Otherwise, if the user runs us as root, and gives us paths that are
> > > > > > only accessible by root, we'll fail to open them, which might in turn
> > > > > > encourage users to change permissions or ownerships: definitely a bad
> > > > > > idea in terms of security.
> > > > > >
> > > > > > Reported-by: Minxi Hou <mhou@redhat.com>
> > > > > > Reported-by: Richard W.M. Jones <rjones@redhat.com>
> > > > > > Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
> > > > >
> > > > > Looking at this I did notice a pre-existing, well, maybe not bug
> > > > > exactly, but possibly surprising behaviour, which makes me a
> > > > > bit more nervous now that we can invoke it as root.
> > > > >
> > > > > tap_sock_unix_open() will silently truncate the given socket path to
> > > > > the maximum length for a Unix socket. Which means we could bind(),
> > > > > but also unlink() a path that's not exactly the same as the one the
> > > > > one the user requested. I don't immediately see a way to exploit
> > > > > that, but it's the sort of thing that makes me nervous. I think we
> > > > > should instead outright fail if the given socket path is too long.
> > > >
> > > > Yes, agreed.
> > > >
> > > > It seems as if the latest passt code still does this. Do you want me
> > > > to open a bug about it?
> > >
> > > Yes, please, that, or a patch :)
> >
> > While I was testing this, I found we do seem to check it:
> >
> > https://passt.top/passt/tree/conf.c#n1446
>
> Oh, I thought David was referring to the loop in tap_sock_unix_open(),
> where we try paths in the form "/tmp/passt_%i.socket". But even there,
> we can't exceed UNIX_PATH_MAX.
Actually I was referring to the memcpy() from sock_path to path in
tap_sock_unix_open(). I hadn't thought to check if we'd previously
verified the length of sock_path. So, yeah we seem to be ok here
(although it's not obvious that's the case from within the code of
tap_sock_unix_open()).
> One minor issue remains, though: in conf(), we refuse paths that are
> longer than UNIX_SOCK_MAX (100). That's the maximum index for the
> "/tmp/passt_%i.socket", it happens to be a sane value, but we should use
> UNIX_PATH_MAX (108) instead. I'll fix it, but wait for David's feedback
> first.
Well, apart from that, yes.
--
David Gibson (he or they) | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you, not the other way
| around.
http://www.ozlabs.org/~dgibson
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
next prev parent reply other threads:[~2024-06-21 1:02 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-22 20:59 [PATCH 0/8] Open socket and PID files as root, before switching Stefano Brivio
2024-05-22 20:59 ` [PATCH 1/8] conf: Don't lecture user about starting us as root Stefano Brivio
2024-05-23 1:45 ` David Gibson
2024-05-23 9:52 ` Richard W.M. Jones
2024-05-22 20:59 ` [PATCH 2/8] tap: Move all-ones initialisation of mac_guest to tap_sock_init() Stefano Brivio
2024-05-23 1:46 ` David Gibson
2024-05-23 9:59 ` Richard W.M. Jones
2024-05-23 10:03 ` Richard W.M. Jones
2024-05-22 20:59 ` [PATCH 3/8] passt, tap: Don't use -1 as uninitialised value for fd_tap_listen Stefano Brivio
2024-05-23 1:48 ` David Gibson
2024-05-22 20:59 ` [PATCH 4/8] tap: Split tap_sock_unix_init() into opening and listening parts Stefano Brivio
2024-05-23 10:05 ` Richard W.M. Jones
2024-05-28 7:01 ` David Gibson
2024-05-22 20:59 ` [PATCH 5/8] util: Rename write_pidfile() to pidfile_write() Stefano Brivio
2024-05-23 10:06 ` Richard W.M. Jones
2024-05-22 20:59 ` [PATCH 6/8] passt, util: Move opening of PID file to its own function Stefano Brivio
2024-05-23 10:06 ` Richard W.M. Jones
2024-05-28 7:04 ` David Gibson
2024-05-22 20:59 ` [PATCH 7/8] conf, passt, tap: Open socket and PID files before switching UID/GID Stefano Brivio
2024-05-23 10:10 ` Richard W.M. Jones
2024-05-29 2:35 ` David Gibson
2024-06-20 11:30 ` Richard W.M. Jones
2024-06-20 12:12 ` Stefano Brivio
2024-06-20 12:47 ` Richard W.M. Jones
2024-06-20 14:22 ` Stefano Brivio
2024-06-21 1:02 ` David Gibson [this message]
2024-05-22 20:59 ` [PATCH 8/8] conf, passt.h: Rename pid_file in struct ctx to pidfile Stefano Brivio
2024-05-23 10:11 ` Richard W.M. Jones
2024-05-28 7:07 ` David Gibson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ZnTRLtW8uQ3FefgT@zatzit \
--to=david@gibson.dropbear.id.au \
--cc=mhou@redhat.com \
--cc=passt-dev@passt.top \
--cc=rjones@redhat.com \
--cc=sbrivio@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://passt.top/passt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for IMAP folder(s).