If SRTP sessions exist, don't create new ones when applying answer.

If SRTP sessions exist, don't create new ones when applying answer.

Instead, call the "Update" methods of SrtpSession, which will just call
srtp_update, instead of wiping out the session state completely.

This was causing decryption to stop working when subsequent
offers/answers are applied. We don't know enough about SRTP to
understand the root cause, and I wasn't able to write an integration
test that reproduces the issue... But at least this fixes the bug that
can be reproduced reliably using Hangouts.

BUG=webrtc:8251
Review-Url: https://codereview.webrtc.org/3019443002
Cr-Commit-Position: refs/heads/master@{#19874}
Committed: https://webrtc.googlesource.com/src/+/5ada7acf603e90e71990e9d4ff8f49389f24958c

[Taylor Brandstetter, 2017-09-15 22:20:37]

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

[Taylor Brandstetter, 2017-09-15 22:20:38]

PTAL

[Zhi Huang, 2017-09-15 22:27:20]

On 2017/09/15 22:20:38, Taylor Brandstetter wrote:
> PTAL

lgtm. Thanks for making CL.

[Taylor Brandstetter, 2017-09-15 23:59:42]

The CQ bit was checked by deadbeef@webrtc.org

[commit-bot: I haz the power, 2017-09-15 23:59:45]

CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.webrtc.org/...

[commit-bot: I haz the power, 2017-09-16 00:52:36]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1505519982881150, "parent_rev":
"1c5e6d0a3f6f7ca2f20bbce354fdf4ee92b2ce93", "commit_rev":
"5ada7acf603e90e71990e9d4ff8f49389f24958c"}

[commit-bot: I haz the power, 2017-09-16 00:52:40]

Description was changed from

==========
If SRTP sessions exist, don't create new ones when applying answer.

Instead, call the "Update" methods of SrtpSession, which will just call
srtp_update, instead of wiping out the session state completely.

This was causing decryption to stop working when subsequent
offers/answers are applied. We don't know enough about SRTP to
understand the root cause, and I wasn't able to write an integration
test that reproduces the issue... But at least this fixes the bug that
can be reproduced reliably using Hangouts.

BUG=webrtc:8251
==========

to

==========
If SRTP sessions exist, don't create new ones when applying answer.

Instead, call the "Update" methods of SrtpSession, which will just call
srtp_update, instead of wiping out the session state completely.

This was causing decryption to stop working when subsequent
offers/answers are applied. We don't know enough about SRTP to
understand the root cause, and I wasn't able to write an integration
test that reproduces the issue... But at least this fixes the bug that
can be reproduced reliably using Hangouts.

BUG=webrtc:8251

Review-Url: https://codereview.webrtc.org/3019443002
Cr-Commit-Position: refs/heads/master@{#19874}
Committed:
https://webrtc.googlesource.com/src/+/5ada7acf603e90e71990e9d4ff8f49389f24958c
==========

[commit-bot: I haz the power, 2017-09-16 00:52:41]

Committed patchset #1 (id:1) as
https://webrtc.googlesource.com/src/+/5ada7acf603e90e71990e9d4ff8f49389f24958c

[henrika_webrtc, 2017-09-18 09:16:54]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.webrtc.org/3017543002/ by henrika@webrtc.org.

The reason for reverting is: Speculative revert since all Android bots on WebRTC
FYI started to fail when this CL landed.

https://build.chromium.org/p/chromium.webrtc.fyi/builders/Android%20Tests%20%....

[phoglund, 2017-09-18 09:26:04]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.webrtc.org/3012393002/ by phoglund@webrtc.org.

The reason for reverting is: Looks like it causes a crash on Chrome Android
browser tests:
https://build.chromium.org/p/chromium.webrtc.fyi/waterfall?builder=Android%20....