1226437 - Meta CSP should not overwrite referrer policy when speculatively applied

Status: RESOLVED FIXED

(Core :: DOM: Security, defect)

Comment 1

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

We should treat referrer policies found in a <meta csp> the same way as we treat <meta referrer>. There's a bunch of code in place to handle that which makes sure that we never use a resource fetched with the wrong referrer.

Comment 2

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

...so it'd be nice to leverage that existing code.

Comment 3

[Christoph Kerschbaumer [:ckerschb]]

Attachment: bug_1226437_meta_csp_referrer_regression.patch

Jonas, as discussed over IRC speculate meta csp should only set speculative referrer policy and not the *actual* referrer policy.

Comment 4

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Comment on attachment 8689848 [details] [diff] [review]
bug_1226437_meta_csp_referrer_regression.patch

Review of attachment 8689848 [details] [diff] [review]:
-----------------------------------------------------------------

::: parser/html/nsHtml5TreeOpExecutor.cpp
@@ +1022,5 @@
> +{
> +  // Record "speculated" referrer policy locally and thread through the
> +  // speculation phase.  The actual referrer policy will be set by
> +  // HTMLMetaElement::BindToTree().
> +  mSpeculationReferrerPolicy = aReferrerPolicy;

Don't move this function. It'll just make hg-history messier.

Comment 5

[Christoph Kerschbaumer [:ckerschb]]

Attachment: bug_1226437_meta_csp_referrer_regression.patch

Comment 6

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/0db00ad2b8a8

Comment 7

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/0db00ad2b8a8