1022343 - Use the hash function's block size as the default length when generating keys

Status: RESOLVED FIXED

(Core :: DOM: Security, defect)

Description

[Tim Taubert [:ttaubert] (inactive)]

Bug 1020882 added default values for key lengths when generating HMAC keys. Unfortunately I got that wrong and used the hash's output size instead of its block size. Easy to fix.

Comment 1

[Tim Taubert [:ttaubert] (inactive)]

Attachment: 0001-Bug-1022343-Use-the-hash-function-s-block-size-as-th.patch

Comment 2

[Richard Barnes [:rbarnes]]

Comment on attachment 8436480 [details] [diff] [review]
0001-Bug-1022343-Use-the-hash-function-s-block-size-as-th.patch

Review of attachment 8436480 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/crypto/WebCryptoTask.cpp
@@ +1038,5 @@
>  
>        if (params.mLength.WasPassed()) {
>          mLength = params.mLength.Value();
>        } else {
>          KeyAlgorithm hashAlg(global, hashName);

r+ for now so that we get the lengths right.  But please replace this with MapAlgorithmNameToMechanism as soon as that's available.

Comment 3

[Tim Taubert [:ttaubert] (inactive)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/b33e1b8f2112

Comment 4

[Tim Taubert [:ttaubert] (inactive)]

Richard, do you think we should request an Aurora uplift given that we ship with wrong default key lengths in Firefox 32? There probably are no consumers so it shouldn't really be a big deal yet.

Comment 5

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/b33e1b8f2112

Comment 6

[Richard Barnes [:rbarnes]]

I don't really think this is critical enough to uplift.  I'm not really seeing any usage in telemetry.