978944 - Use Android native code pbkdf2sha256 (from NativeCrypto library) in FxAccountClient

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: Android Sync, defect, P2)

Description

[Nick Alexander :nalexander [he/him]]

For great perf win.  It might be wise to fallback to our Java code in the case when we can't load the library, but I'm not particularly concerned.

Comment 1

[Nick Alexander :nalexander [he/him]]

The relevant helper is

http://mxr.mozilla.org/mozilla-central/source/mobile/android/base/background/fxa/FxAccountUtils.java#107

and it's invoked from

http://mxr.mozilla.org/mozilla-central/source/mobile/android/base/background/fxa/QuickPasswordStretcher.java#28

Testing this across devices is more difficult than the fix, for sure.

Comment 2

[Richard Newman [:rnewman]]

Also measure the win!

Comment 3

[Nick Alexander :nalexander [he/him]]

Attachment: github PR

mcomella: feel free to steal this r?.

As for quantifying the win: we go from seconds to milliseconds on my Samsung Galaxy SII.  I had to check that we were, in fact, showing the spinner at all.

Comment 4

[Michael Comella (:mcomella) [NI reported issues only: ex-Mozilla]]

Comment on attachment 8390755 [details] [review]
github PR

r+ w/ nits.

Comment 5

[Nick Alexander :nalexander [he/him]]

https://hg.mozilla.org/integration/fx-team/rev/f5109b04e083

Comment 6

[Nick Alexander :nalexander [he/him]]

(In reply to Nick Alexander :nalexander from comment #5)
> https://hg.mozilla.org/integration/fx-team/rev/f5109b04e083

Damn, wrong URL.  How is this not automated yet?

https://hg.mozilla.org/integration/fx-team/rev/fb1c7f3a3a77

Comment 7

[Mark Finkle (:mfinkle) (use needinfo?)]

(In reply to Nick Alexander :nalexander from comment #3)

> As for quantifying the win: we go from seconds to milliseconds on my Samsung
> Galaxy SII.  I had to check that we were, in fact, showing the spinner at
> all.

\o/

Comment 8

[Richard Newman [:rnewman]]

Posted in Bug 915312 inquiring about uplifting this pair. If it knocks down our memory footprint and has that dramatic an effect on the first-time experience for a user, IMO we should consider getting it out in 29 (and maybe the 400msec startup win in Bug 959652, too).

If we made a mistake, we have ~7 weeks to change our minds.

Comment 9

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/fb1c7f3a3a77

Comment 10

[Nick Alexander :nalexander [he/him]]

(In reply to Richard Newman [:rnewman] from comment #8)
> Posted in Bug 915312 inquiring about uplifting this pair. If it knocks down
> our memory footprint and has that dramatic an effect on the first-time
> experience for a user, IMO we should consider getting it out in 29 (and
> maybe the 400msec startup win in Bug 959652, too).

There were two things at play in the perf wins: this code *and* a server misconfiguration that caused the remote request to be significantly faster.  I still think we get significant perf wins from this code, just not "blink of an eye".

In any case, I think we should uplift if Bug 915312 lands cleanly on Aurora.

Comment 11

[Michael Comella (:mcomella) [NI reported issues only: ex-Mozilla]]

Bug 915312 has been uplifted to Aurora (29).