Tune two SSL Server Socket parameters.

Tune two SSL Server Socket parameters.

Reduce the maximum number of Session ID cache entries to save
memory but increase the session ID cache timeout to be useful.

BUG=426630
Committed: https://crrev.com/91f5db33db20e0aece28f0b66af0f6f53df5fd4c
Cr-Commit-Position: refs/heads/master@{#302373}

[GusFernandez, 2014-10-29 20:37:34]

gusfernandez@chromium.org changed reviewers:
+ davidben@chromium.org, dougsteed@chromium.org, lcwu@chromium.org,
rsleevi@chromium.org

[GusFernandez, 2014-10-29 20:38:49]

On 2014/10/29 20:37:34, gusfernandez wrote:
> mailto:gusfernandez@chromium.org changed reviewers:
> + mailto:davidben@chromium.org, mailto:dougsteed@chromium.org,
mailto:lcwu@chromium.org,
> mailto:rsleevi@chromium.org

This is an alternative to https://codereview.chromium.org/679473002 as discussed
previously.

[GusFernandez, 2014-10-29 20:39:20]

An alternative approach to https://codereview.chromium.org/679473002/

[Ryan Sleevi, 2014-10-29 20:55:24]

rsleevi@chromium.org changed reviewers:
+ juberti@chromium.org, sergeyu@chromium.org

[Ryan Sleevi, 2014-10-29 20:55:24]

Adding Sergey (Remoting, p2p socket) and Justin (WebRTC) to comment on whether
these numbers are acceptable for them.

TL;DR: 64 TLS sessions cached max, process wide (if a client doesn't support TLS
session tickets). Does this affect your APIs?

[Ryan Sleevi, 2014-10-29 20:55:56]

rsleevi@chromium.org changed reviewers:
- sergeyu@chromium.org

[Ryan Sleevi, 2014-10-29 20:56:46]

rsleevi@chromium.org changed reviewers:
+ jamiewalch@chromium.org, mallinath@chromium.org

[Ryan Sleevi, 2014-10-29 20:56:47]

-Sergey (who is OOO)
+jamie for the chromoting question
+mallinath for the P2P sockets question

[Mallinath (Gone from Chromium), 2014-10-30 17:43:18]

mallinath@chromium.org changed reviewers:
+ jiayl@chromium.org
- mallinath@chromium.org

[Mallinath (Gone from Chromium), 2014-10-30 17:43:39]

On 2014/10/29 20:56:47, Ryan Sleevi wrote:
> -Sergey (who is OOO)
> +jamie for the chromoting question
> +mallinath for the P2P sockets question

-mallinath, +jiayl for P2P socket

[Wez, 2014-10-30 19:05:14]

Remoting doesn't re-use SSL sessions, so I don't think this change affects us.

[jiayl, 2014-10-30 20:19:01]

On 2014/10/30 17:43:39, mallinath2 wrote:
> On 2014/10/29 20:56:47, Ryan Sleevi wrote:
> > -Sergey (who is OOO)
> > +jamie for the chromoting question
> > +mallinath for the P2P sockets question
> 
> -mallinath, +jiayl for P2P socket

I don't find any p2p socket or webrtc code using these classes.

[davidben, 2014-10-31 23:31:14]

lgtm

[juberti2, 2014-10-31 23:42:10]

On 2014/10/31 23:31:14, David Benjamin wrote:
> lgtm

According to the code in
https://code.google.com/p/webrtc/source/browse/trunk/webrtc/base/nssstreamada...,
we
- disable caching
- disable session tickets
- use SSL_ConfigSecureServer and SSL_ImportFD to pass in our own 'socket' and
configure SSL

so while I'm not sure whether this change affects us, it's not going to make
anything worse.

[davidben, 2014-10-31 23:55:35]

On 2014/10/31 23:42:10, juberti2 wrote:
> On 2014/10/31 23:31:14, David Benjamin wrote:
> > lgtm
> 
> According to the code in
>
https://code.google.com/p/webrtc/source/browse/trunk/webrtc/base/nssstreamada...,
> we
> - disable caching
> - disable session tickets
> - use SSL_ConfigSecureServer and SSL_ImportFD to pass in our own 'socket' and
> configure SSL
> 
> so while I'm not sure whether this change affects us, it's not going to make
> anything worse.

Yeah, if anything, it'll probably make you use less memory. :-) But you don't
seem to enable //net's SSL server socket, so I think it's a no-op. (Though,
given that NSS's session cache seems to be made entirely of globals, I certainly
hope none of these uses of NSS's server bits ever run in the same process...)

[GusFernandez, 2014-11-01 00:26:26]

The CQ bit was checked by gusfernandez@chromium.org

[commit-bot: I haz the power, 2014-11-01 00:26:46]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/687343002/1

[commit-bot: I haz the power, 2014-11-01 01:08:16]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2014-11-01 01:08:54]

Patchset 1 (id:??) landed as
https://crrev.com/91f5db33db20e0aece28f0b66af0f6f53df5fd4c
Cr-Commit-Position: refs/heads/master@{#302373}