Message-ID: <92049e63-217f-4122-b84a-401d0f7c214a@...1227...> Date: Wed, 31 Jul 2024 00:59:48 +0200 MIME-Version: 1.0 References: <20240730170021.1101321-1-storehouse@...2805...> From: Arne Schwabe In-Reply-To: <20240730170021.1101321-1-storehouse@...2805...> Subject: Re: [Openvpn-devel] [PATCH] examples: Switched to ed25519, nodes -> noenc List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="utf-8"; format="flowed" Content-Transfer-Encoding: 8bit To: Eric Toombs , openvpn-devel@lists.sourceforge.net Am 30.07.2024 um 19:00 schrieb Eric Toombs via Openvpn-devel: > From: Eric Toombs We prefer real email addresses in our commits. > > I switched the curve to ed25519, a generally more trustworthy curve and the default in openssh. > As a bonus, it *really* simplifies what is already a complicated command. > The old command wouldn't even run in all shells because it used process substitution. > > 'nodes' is deprecated in favour of 'noenc', so I switched that too. While ed25519 might have its advantages, its usage in X509 certificates is really really uncommon. I have seen more brainpoool usages in the wild than ed25519. It is also not allowed on hosts that use FIPS settings and also older OpenVPN clients/OpenSSL clients might not support it. Also in contrast to secp384r1, I know absolutely no secure certficate store that supports it (Android Keystore, YubiKey and so on all support secp384r1 but there is no ed25519 support).  So I don't think we are at the point where we want to recommend in an easy-to-setup guide a group/curve that is a lot less supported to the basically universially supported secp384r1. So for the purpose of this guide, I think secp384r1 is a much better choice compared to ed25519. Arne