Date: Tue, 22 Feb 2022 13:38:27 +0100 (CET) From: Frank Lichtenheld Message-ID: <828192399.42381.1645533507517@...2642...> In-Reply-To: <7327aea6-5cd7-7b67-b92d-30d3da50a53f@...1227...> References: <20220221111933.1314-1-frank@...2641...> <20220222121439.1260-1-frank@...2641...> <7327aea6-5cd7-7b67-b92d-30d3da50a53f@...1227...> MIME-Version: 1.0 Subject: Re: [Openvpn-devel] [PATCH v2] doc: cleanup for --data-ciphers and related List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit To: Arne Schwabe , openvpn-devel@lists.sourceforge.net > Arne Schwabe hat am 22.02.2022 13:21 geschrieben: > > > > > > The ``algorithm`` parameter may be :code:`lzo`, :code:`lz4`, > > @@ -193,6 +193,10 @@ configured in a compatible way between both the local and remote side. > > supported by the client will be pushed to clients that support cipher > > negotiation. > > > > + For more details see the chapter on `Data channel cipher negotiation`_. > > + *Especially* if you need to support clients with OpenVPN versions older > > + than 2.5! > > + > > Why 2.5? 2.4 client and server will just default to AES-256-GCM and that > should be fine? I used < 2.5 because I removed the explanation of the 2.4 quirk from below. > > - > > - Note for using NCP with an OpenVPN 2.4 peer: This list must include the > > - :code:`AES-256-GCM` and :code:`AES-128-GCM` ciphers. > > + If ``--compat-mode`` is set to a version older than 2.5.0 ``--cipher`` > > + will be appended to ``--data-ciphers`` if not already present. > > Nitpick, not ``--cipher`` but the cipher specified by ``--cipher`` Okay. > > @@ -3219,11 +3219,11 @@ options_set_backwards_compatible_options(struct options *o) > > } > > > > /* Versions < 2.5.0 do need --cipher in the list of accepted ciphers. > > - * Version 2.4 might probably does not need it but NCP was not so > > + * Version 2.4 probably does not need it but NCP was not so > > * good with 2.4 and ncp-disable might be more common on 2.4 peers. > > - * Only do this iif --cipher is not explicitly (BF-CBC). This is not > > - * 100% correct backwards compatible behaviour but 2.5 already behaved like > > - * this */ > > + * Only do this if --cipher is set explicitly (or compat mode is > > iif should have been iff (=if and only if). So probably either correct > to that or write it out if you think that math idiom is not understood > by normal programmers. Yeah, I guessed that. But I thought it weird to use iff and then immediately add in parathesis an exception to that statement. So I will leave it like this. > > + * < 2.4.0, see above). This is not 100% correct backwards compatible > > + * behaviour but 2.5 already behaved like this */ > > if (o->ciphername && need_compatibility_before(o, 20500) > > && !tls_item_in_cipher_list(o->ciphername, o->ncp_ciphers)) > > { -- Frank Lichtenheld