1215796 - Cleanup some non-secure fallback options

Status: RESOLVED FIXED

(Core :: Security: PSM, defect)

Description

[Masatoshi Kimura [:emk]]

Once we have an override UX, we can remove some redundant options.

Comment 1

[Masatoshi Kimura [:emk]]

Attachment: Remove unrestricted RC4 fallback pref

This is an easy footgun.

Comment 2

[Masatoshi Kimura [:emk]]

Attachment: Remove the static fallback whitelist

Unrestricted RC4 fallback was enabled until Firefox 43. The static fallback will be disabled since Firefox 44. So this list is just a waste of binary size.

Comment 3

[Masatoshi Kimura [:emk]]

Attachment: Bump the lowest valid fallback limit to 2 (TLS 1.1)

Chrome 45+ already disabled the non-secure fallback to TLS 1.0. Chrome 48+ will remove the option to re-enable the fallback.

Comment 4

[Masatoshi Kimura [:emk]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=7fc65622f2b6

Comment 5

[Dana Keeler (she/her) (use needinfo) [:keeler]]

Comment on attachment 8675339 [details] [diff] [review]
Remove the static fallback whitelist

Review of attachment 8675339 [details] [diff] [review]:
-----------------------------------------------------------------

Great

Comment 6

[Dana Keeler (she/her) (use needinfo) [:keeler]]

Comment on attachment 8675338 [details] [diff] [review]
Remove unrestricted RC4 fallback pref

Review of attachment 8675338 [details] [diff] [review]:
-----------------------------------------------------------------

I'm concerned about removing this so quickly after shipping the new UI. If we find a significant problem, we're going to want an easy, low-risk change we can make to restore the original behavior (as in, flipping a pref). Let's keep this for a release or two until we're confident about it (for one thing, no RC4 override UI has been implemented for android or b2g, as far as I'm aware).

Comment 7

[Dana Keeler (she/her) (use needinfo) [:keeler]]

Comment on attachment 8675340 [details] [diff] [review]
Bump the lowest valid fallback limit to 2 (TLS 1.1)

Review of attachment 8675340 [details] [diff] [review]:
-----------------------------------------------------------------

I seem to recall some aspect of RC4 fallback requiring TLS 1.0. Is this no longer the case?
Anyway, I don't see this as a change we urgently need to make, seeing as the pref defaults to a fallback limit of TLS 1.2.

Comment 8

[Masatoshi Kimura [:emk]]

(In reply to David Keeler [:keeler] (use needinfo?) from comment #7)
> I seem to recall some aspect of RC4 fallback requiring TLS 1.0. Is this no
> longer the case?

IE requires to fallback to TLS 1.0, but we don't. We had to support RC4 fallback with TLS 1.2 because of FALLBACK_SCSV.

Comment 9

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/e7e994b6a5a3

Comment 10

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

https://hg.mozilla.org/mozilla-central/rev/e7e994b6a5a3

Comment 11

[Masatoshi Kimura [:emk]]

Ah, forgot to put a leave-open.

Comment 12

[Dana Keeler (she/her) (use needinfo) [:keeler]]

:emk, since we've gone through a few release cycles since the first patch in this bug landed, it would be best to land any remaining patches in a separate bug (this makes it easier for everyone to track what landed in which version).

Comment 13

[Jean-Yves Perrier [:teoli]]

Added a note in https://developer.mozilla.org/en-US/Firefox/Releases/44#Security