- From: Henri Sivonen <hsivonen@hsivonen.fi>
- Date: Mon, 12 Jan 2015 15:52:31 +0200
- To: Marc Fawzi <marc.fawzi@gmail.com>
- Cc: David Sheets <kosmo.zb@gmail.com>, "Eric J. Bowman" <eric@bisonsystems.net>, Domenic Denicola <d@domenic.me>, Tim Berners-Lee <timbl@w3.org>, Chris Palmer <palmer@google.com>, Melvin Carvalho <melvincarvalho@gmail.com>, Mark Nottingham <mnot@mnot.net>, Public TAG List <www-tag@w3.org>
(Reordering quotes) >> On Jan 9, 2015, at 9:33 AM, David Sheets <kosmo.zb@gmail.com> wrote: >> Web Crypto over http + SRI with digests delivered over a certified >> transport (e.g. TLS) should be secure. On Fri, Jan 9, 2015 at 11:06 PM, Marc Fawzi <marc.fawzi@gmail.com> wrote: > Where does this put the Chrome's team assertion that https is needed for Web Crypto? It leaves the assertion as being correct. 1) David's answer assumes that you'd be able to waive mixed-content blocking using SRI, but that's not a feature that exists in present tense (AFAIK). Also, it's an open issue in the current draft http://w3c.github.io/webappsec/specs/subresourceintegrity/ 2) Even if we assume that SRI will allow you to waive confidentiality of subresources, the integrity still chains to the main resource that needs to be delivered via https for Web Crypto to be secure. So even if SRI ends up allowing the waiver of confidentiality for subresources, the integrity of the JS code running under the authority of the origin purporting to do something secure with Web Crypto still needs to chain to https for there to be protection against active MITMs. What made you read David's words "certified transport (e.g. TLS)" as not meaning https? > Don't expect you to be able to answer all of these questions but I feel like Google needs to evaluate whether they have the right expertise in security on the Chrome team or if the web's most popular browser is driven by false assumptions in this area. They have the right expertise. -- Henri Sivonen hsivonen@hsivonen.fi https://hsivonen.fi/
Received on Monday, 12 January 2015 13:52:58 UTC