Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 10 Feb 2023 06:23:05 +0800
From:      "Ben Woods" <woodsb02@freebsd.org>
To:        freebsd-security@freebsd.org
Cc:        "Nathan Dorfman" <ndorf@rtfm.net>, "Mariusz Zaborski" <oshogbo@FreeBSD.org>, "Gordon Tetlow" <gordon@FreeBSD.org>, "Philip Paeps" <philip@freebsd.org>, "Alan Somers" <asomers@freebsd.org>, "Maksym Sobolyev" <sobomax@freebsd.org>
Subject:   Re: FreeBSD Security Advisory FreeBSD-SA-23:01.geli
Message-ID:  <b2994552-139d-4b11-b459-2d1fa087f183@app.fastmail.com>
In-Reply-To: <20230208190833.283D087C3@freefall.freebsd.org>
References:  <20230208190833.283D087C3@freefall.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 9 Feb 2023, at 3:08 AM, FreeBSD Security Advisories wrote:
> FreeBSD-SA-23:01.geli                                       Security A=
dvisory
>                                                           The FreeBSD =
Project
>
> Topic:          GELI silently omits the keyfile if read from stdin

Good morning,

I was scrolling through my emails yesterday and spat my coffee out when =
I read this one. I just wanted to put my hand up and say I believe this =
issue originates from my code, when I added the =E2=80=9Cgeli init multi=
ple providers=E2=80=9D feature in 2018 just prior to the FreeBSD-12 rele=
ase.

https://reviews.freebsd.org/D16115
https://reviews.freebsd.org/D17096

Apologies to anyone affected, and thank you to Nathan for reporting it, =
Marius, Gordon and Philip for fixing it, and anyone else on the security=
 team for investigating/communicating the issue.

I=E2=80=99ll spend some time to review the fix to fully understand where=
 I went wrong. I was also wondering why it wasn=E2=80=99t revealed by my=
 testing at the time=E2=80=A6. And then I realised this would not be vis=
ible to the user as they would still enter their user key to successfull=
y add the device with a null master key. Slaps forehead.

I never got around to adding unit tests for init/attach multiple provide=
rs as was requested by Alan Somers at the time (sorry), but I suspect ev=
en if I had they would have passed because I wouldn=E2=80=99t have thoug=
ht to test for this scenario.

Regards,
Ben

--=20
From: Ben Woods
woodsb02@freebsd.org



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?b2994552-139d-4b11-b459-2d1fa087f183>