Stop preemptively generating an RSA key pair.

Stop preemptively generating an RSA key pair.

RSA isn't used by the Java binding, it isn't the default for Obj-C,
and this identity store class isn't used at all by Chrome. So in most
cases, preemptively generating an RSA key pair just wastes CPU cycles
and blocks the worker thread.

If a native C++ application really wants to preemptively generate an RSA
key pair, it can easily do this by passing in its own DtlsIdentityStoreImpl
and calling GenerateIdentity on it.

R=juberti@chromium.org

Committed: https://crrev.com/9cb23a3579990ff78e3ceb6b6a88a06912662d7b
Cr-Commit-Position: refs/heads/master@{#12498}

[Taylor Brandstetter, 2016-04-22 20:56:16]

deadbeef@webrtc.org changed reviewers:
+ pthatcher@webrtc.org

[Taylor Brandstetter, 2016-04-22 20:56:16]

PTAL Peter.

FYI Justin, this is actually what was causing the delay I mentioned in our
standup today. Even if an application uses ECDSA, an extra RSA key pair that's
never used ends up being generated on the worker thread.

[juberti2, 2016-04-23 06:06:06]

juberti@chromium.org changed reviewers:
+ juberti@chromium.org

[juberti2, 2016-04-23 06:06:06]

lgtm. 

What's the cost to generate an ECDSA cert on mobile? Would we ever want to do
that pre-emptively?

[Taylor Brandstetter, 2016-04-25 17:16:32]

On 2016/04/23 06:06:06, juberti2 wrote:
> lgtm. 
> 
> What's the cost to generate an ECDSA cert on mobile? Would we ever want to do
> that pre-emptively?

Very, very fast in comparison (a few milliseconds). I don't think it's worth
generating pre-emptively.

[Taylor Brandstetter, 2016-04-25 19:50:06]

The CQ bit was checked by deadbeef@webrtc.org

[commit-bot: I haz the power, 2016-04-25 19:50:13]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1907083005/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1907083005/1

[commit-bot: I haz the power, 2016-04-25 20:03:10]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-04-25 20:03:11]

Try jobs failed on following builders:
  mac_rel on tryserver.webrtc (JOB_FAILED,
http://build.chromium.org/p/tryserver.webrtc/builders/mac_rel/builds/14575)

[Taylor Brandstetter, 2016-04-25 20:50:50]

Justin, can you lgtm from @webrtc.org account?

[juberti1, 2016-04-26 00:36:02]

On 2016/04/25 20:50:50, Taylor Brandstetter wrote:
> Justin, can you lgtm from @webrtc.org account?

lgtm

[Taylor Brandstetter, 2016-04-26 00:39:40]

Description was changed from

==========
Stop preemptively generating an RSA key pair.

RSA isn't used by the Java binding, it isn't the default for Obj-C,
and this identity store class isn't used at all by Chrome. So in most
cases, preemptively generating an RSA key pair just wastes CPU cycles
and blocks the worker thread.

If a native C++ application really wants to preemptively generate an RSA
key pair, it can easily do this by passing in its own DtlsIdentityStoreImpl
and calling GenerateIdentity on it.
==========

to

==========
Stop preemptively generating an RSA key pair.

RSA isn't used by the Java binding, it isn't the default for Obj-C,
and this identity store class isn't used at all by Chrome. So in most
cases, preemptively generating an RSA key pair just wastes CPU cycles
and blocks the worker thread.

If a native C++ application really wants to preemptively generate an RSA
key pair, it can easily do this by passing in its own DtlsIdentityStoreImpl
and calling GenerateIdentity on it.

R=juberti@chromium.org

Committed:
https://chromium.googlesource.com/external/webrtc/+/9cb23a3579990ff78e3ceb6b6...
==========

[Taylor Brandstetter, 2016-04-26 00:39:41]

Committed patchset #2 (id:20001) manually as
9cb23a3579990ff78e3ceb6b6a88a06912662d7b (presubmit successful).

[commit-bot: I haz the power, 2016-04-26 00:39:45]

Description was changed from

==========
Stop preemptively generating an RSA key pair.

RSA isn't used by the Java binding, it isn't the default for Obj-C,
and this identity store class isn't used at all by Chrome. So in most
cases, preemptively generating an RSA key pair just wastes CPU cycles
and blocks the worker thread.

If a native C++ application really wants to preemptively generate an RSA
key pair, it can easily do this by passing in its own DtlsIdentityStoreImpl
and calling GenerateIdentity on it.

R=juberti@chromium.org

Committed:
https://chromium.googlesource.com/external/webrtc/+/9cb23a3579990ff78e3ceb6b6...
==========

to

==========
Stop preemptively generating an RSA key pair.

RSA isn't used by the Java binding, it isn't the default for Obj-C,
and this identity store class isn't used at all by Chrome. So in most
cases, preemptively generating an RSA key pair just wastes CPU cycles
and blocks the worker thread.

If a native C++ application really wants to preemptively generate an RSA
key pair, it can easily do this by passing in its own DtlsIdentityStoreImpl
and calling GenerateIdentity on it.

R=juberti@chromium.org

Committed: https://crrev.com/9cb23a3579990ff78e3ceb6b6a88a06912662d7b
Cr-Commit-Position: refs/heads/master@{#12498}
==========

[commit-bot: I haz the power, 2016-04-26 00:39:46]

Patchset 2 (id:??) landed as
https://crrev.com/9cb23a3579990ff78e3ceb6b6a88a06912662d7b
Cr-Commit-Position: refs/heads/master@{#12498}

[pthatcher1, 2016-04-29 22:42:01]

lgtm