Re: [TLS] Cipher suite values to indicate TLS capability

Adam Langley <agl@google.com> Tue, 05 June 2012 23:51 UTC

MIME-Version: 1.0
In-Reply-To: <m2oboxxr5z.fsf@localhost.localdomain>
References: <CAL9PXLwdQctUub5oPx0tepsfveDo0bNKGBUaUBBFeq4u4D0BbA@mail.gmail.com> <m2sje9xsc0.fsf@localhost.localdomain> <CAL9PXLy_Lr+-ehOKSddtooVBpgUzxCyLKhWghC7UtOAt3HH2Rw@mail.gmail.com> <m2oboxxr5z.fsf@localhost.localdomain>
Date: Tue, 05 Jun 2012 19:51:47 -0400
Message-ID: <CAL9PXLx4qdv1AYv7f47=1f8j7UkkOuaYEn9PHeLnQRWkJ5NKDQ@mail.gmail.com>
From: Adam Langley <agl@google.com>
To: Geoffrey Keating <geoffk@geoffk.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Cc: tls@ietf.org
Subject: Re: [TLS] Cipher suite values to indicate TLS capability
Precedence: list

On Tue, Jun 5, 2012 at 7:41 PM, Geoffrey Keating <geoffk@geoffk.org> wrote:
> Why would you need to send the extensions?  (The RFC says they're
> optional.)

Well, we could cross our fingers and hope that the server picks
something acceptable I guess (or simply define the set of acceptable
values for SSLv3). But extending SSLv3 to support ECDHE seems rather
specific:

There's also the issue of captive portals which tend to trigger
fallbacks by doing things like sending HTTP replies on port 443. If
the user authorises during the fallback process then we downgrade to
SSLv3, omit SNI and throw a nasty certificate error for any SNI
requiring domain names. I'd rather throw a network error than a
certificate error that trains users to ignore them. (This has actually
been happening to some of our users, although we significantly
ameliorated it by being more aggressive in stepping back up to TLS
after a fallback.)

Cheers

AGL

Re: [TLS] Cipher suite values to indicate TLS cap… Geoffrey Keating
[TLS] Cipher suite values to indicate TLS capabil… Adam Langley
Re: [TLS] Cipher suite values to indicate TLS cap… Adam Langley
Re: [TLS] Cipher suite values to indicate TLS cap… Geoffrey Keating
Re: [TLS] Cipher suite values to indicate TLS cap… Adam Langley
Re: [TLS] Cipher suite values to indicate TLS cap… Adam Langley
Re: [TLS] Cipher suite values to indicate TLS cap… Geoffrey Keating
Re: [TLS] Cipher suite values to indicate TLS cap… Adam Langley
Re: [TLS] Cipher suite values to indicate TLS cap… Chris Richardson
Re: [TLS] Cipher suite values to indicate TLS cap… Wan-Teh Chang
Re: [TLS] Cipher suite values to indicate TLS cap… Yoav Nir
Re: [TLS] Cipher suite values to indicate TLS cap… Yoav Nir
Re: [TLS] Cipher suite values to indicate TLS cap… Nikos Mavrogiannopoulos
Re: [TLS] Cipher suite values to indicate TLS cap… Geoffrey Keating
Re: [TLS] Cipher suite values to indicate TLS cap… Chris Richardson
Re: [TLS] Cipher suite values to indicate TLS cap… Tom Ritter
Re: [TLS] Cipher suite values to indicate TLS cap… Nikos Mavrogiannopoulos
Re: [TLS] Cipher suite values to indicate TLS cap… Adam Langley
Re: [TLS] Cipher suite values to indicate TLS cap… Nikos Mavrogiannopoulos
Re: [TLS] Cipher suite values to indicate TLS cap… Adam Langley
Re: [TLS] Cipher suite values to indicate TLS cap… Martin Rex