Align SSLClientSocketOpenSSL and SSLClientSocketNSS histograms.

Align SSLClientSocketOpenSSL and SSLClientSocketNSS histograms.

This makes a handful of small changes to align the two implementations:

- Remove Net.SSLv3FallbackToRenegoPatchedServer. It's no longer necessary. In
  fact, the majority of SSLv3 fallbacks are to servers that /are/ renego patched.

- Gather Net.SSLCertVerificationTime* on OpenSSL.

- Move histogram calls on OpenSSL out of GetSSLInfo as that is not called once
  per handshake. Net.RenegotiationExtensionSupported is moved to common code and
  RecordChannelIDSupport to the handshake success codepath.

Net.OCSPResponseStapled is left ungathered as BoringSSL does not yet support
OCSP stapling.

BUG=405631
Committed: https://crrev.com/09c3d07392bc432e5ef16874b5b3cdc5074b1cb4
Cr-Commit-Position: refs/heads/master@{#291752}

[davidben, 2014-08-20 19:47:15]

I can split these up into three separate CLs if you prefer.

[Ryan Sleevi, 2014-08-20 19:58:12]

LGTM, mod one request from Adam below.

AGL should comment on the SSL3 fallback experiment, but I think that's fine to
remove.

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
File net/socket/ssl_client_socket_nss.cc (left):

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
net/socket/ssl_client_socket_nss.cc:2477: peer_supports_renego_ext == PR_TRUE);
Adam should comment on this. I think I'm fine removing this, in light of the
fallback scsv.

[agl, 2014-08-22 18:26:33]

lgtm

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
File net/socket/ssl_client_socket_nss.cc (left):

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
net/socket/ssl_client_socket_nss.cc:2477: peer_supports_renego_ext == PR_TRUE);
On 2014/08/20 19:58:12, Ryan Sleevi wrote:
> Adam should comment on this. I think I'm fine removing this, in light of the
> fallback scsv.

LGTM. I don't that we need to worry about this any more - the numbers suggest
that it's never going to work.

[davidben, 2014-08-22 18:28:51]

Right, I always forget this one... +asvitkine for histograms.xml.

[Alexei Svitkine (slow), 2014-08-22 18:49:14]

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
File net/socket/ssl_client_socket_openssl.cc (right):

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
net/socket/ssl_client_socket_openssl.cc:1023:
UMA_HISTOGRAM_TIMES("Net.SSLCertVerificationTime", verify_time);
Nit: Bad indentation.

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
File net/socket/ssl_client_socket_pool.cc (right):

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
net/socket/ssl_client_socket_pool.cc:508:
UMA_HISTOGRAM_ENUMERATION("Net.RenegotiationExtensionSupported", 1, 2);
Nit: Can you just use an UMA_HISTOGRAM_BOOLEAN() here?

Also, can you refactor to have a single macro invocation for the same histogram?
Each macro expands to a bunch of code, so fewer of them results in less bloat.

[davidben, 2014-08-22 23:03:20]

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
File net/socket/ssl_client_socket_openssl.cc (right):

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
net/socket/ssl_client_socket_openssl.cc:1023:
UMA_HISTOGRAM_TIMES("Net.SSLCertVerificationTime", verify_time);
On 2014/08/22 18:49:14, Alexei Svitkine wrote:
> Nit: Bad indentation.

Done.

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
File net/socket/ssl_client_socket_pool.cc (right):

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
net/socket/ssl_client_socket_pool.cc:508:
UMA_HISTOGRAM_ENUMERATION("Net.RenegotiationExtensionSupported", 1, 2);
On 2014/08/22 18:49:14, Alexei Svitkine wrote:
> Nit: Can you just use an UMA_HISTOGRAM_BOOLEAN() here?
> 
> Also, can you refactor to have a single macro invocation for the same
histogram?
> Each macro expands to a bunch of code, so fewer of them results in less bloat.

Done. Though, that should be flipped back if it isn't safe to flip an existing
0/1 enumeration to a boolean. (Is it? This isn't an new histogram.)

[Alexei Svitkine (slow), 2014-08-24 15:49:20]

lgtm

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
File net/socket/ssl_client_socket_pool.cc (right):

https://codereview.chromium.org/493793003/diff/1/net/socket/ssl_client_socket...
net/socket/ssl_client_socket_pool.cc:508:
UMA_HISTOGRAM_ENUMERATION("Net.RenegotiationExtensionSupported", 1, 2);
On 2014/08/22 23:03:20, David Benjamin wrote:
> On 2014/08/22 18:49:14, Alexei Svitkine wrote:
> > Nit: Can you just use an UMA_HISTOGRAM_BOOLEAN() here?
> > 
> > Also, can you refactor to have a single macro invocation for the same
> histogram?
> > Each macro expands to a bunch of code, so fewer of them results in less
bloat.
> 
> Done. Though, that should be flipped back if it isn't safe to flip an existing
> 0/1 enumeration to a boolean. (Is it? This isn't an new histogram.)

The two are equivalent in terms of what gets sent up.

[davidben, 2014-08-25 19:36:42]

The CQ bit was checked by davidben@chromium.org

[commit-bot: I haz the power, 2014-08-25 19:37:28]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/davidben@chromium.org/493793003/20001

[commit-bot: I haz the power, 2014-08-25 20:34:18]

Committed patchset #2 (20001) as 75502d79c21e38bb4309fe814bd3614ca08cb36f

[commit-bot: I haz the power, 2014-09-10 02:37:15]

Patchset 2 (id:??) landed as
https://crrev.com/09c3d07392bc432e5ef16874b5b3cdc5074b1cb4
Cr-Commit-Position: refs/heads/master@{#291752}