ScopedTrackers to help pin down where IOThread::InitSystemRequestContextOnIOThread is janky

ScopedTrackers to help pin down where IOThread::InitSystemRequestContextOnIOThread is janky.

TBR=pkasting
BUG=454983
Committed: https://crrev.com/3e5bced63850fc86a3ad386abd37b9287e1cfff6
Cr-Commit-Position: refs/heads/master@{#315423}

[michaeln, 2015-02-09 21:35:59]

michaeln@chromium.org changed reviewers:
+ pkasting@chromium.org

[michaeln, 2015-02-09 21:36:15]

ptal

[michaeln, 2015-02-09 22:37:48]

tbr'ing since this is purely mechanical

[michaeln, 2015-02-09 22:37:54]

The CQ bit was checked by michaeln@chromium.org

[commit-bot: I haz the power, 2015-02-09 22:39:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/886933010/40001

[commit-bot: I haz the power, 2015-02-09 22:58:33]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-02-09 22:59:31]

Patchset 3 (id:??) landed as
https://crrev.com/3e5bced63850fc86a3ad386abd37b9287e1cfff6
Cr-Commit-Position: refs/heads/master@{#315423}

[Peter Kasting, 2015-02-09 23:35:44]

https://codereview.chromium.org/886933010/diff/40001/chrome/browser/io_thread.cc
File chrome/browser/io_thread.cc (right):

https://codereview.chromium.org/886933010/diff/40001/chrome/browser/io_thread...
chrome/browser/io_thread.cc:1106: 
Do you want a tracker here to measure the remaining time in this function
besides what you're instrumenting below and elsewhere?

In particular it seems like someone suggested ConstructSystemRequestContext()
below might be slow?  Didn't know if you wanted to also track that.

https://codereview.chromium.org/886933010/diff/40001/chrome/browser/io_thread...
chrome/browser/io_thread.cc:1128: ".HttpNetworkSession"));
Nit: Name seems slightly misleading since this is actually measuring the
HttpNetworkLayer/HttpTransactionFactory constructor costs, right?  (Since you've
instrumented HttpNetworkSession's constructor separately.)

[michaeln, 2015-02-10 00:02:24]

https://codereview.chromium.org/886933010/diff/40001/chrome/browser/io_thread.cc
File chrome/browser/io_thread.cc (right):

https://codereview.chromium.org/886933010/diff/40001/chrome/browser/io_thread...
chrome/browser/io_thread.cc:1106: 
On 2015/02/09 23:35:44, Peter Kasting wrote:
> Do you want a tracker here to measure the remaining time in this function
> besides what you're instrumenting below and elsewhere?

No, i'm trying to distinguish between objs being constructed in
HttpNetworkSession's initializer list vs the body of that ctor.
 
> In particular it seems like someone suggested ConstructSystemRequestContext()
> below might be slow?  Didn't know if you wanted to also track that.

ConstructSystemRequestContext is tracked sepereately (see line 258)

I'm wondering about CertDatabase, it's platform specific and the overall
InitSystemRequestContext slowness is specific to macs.

[michaeln, 2015-02-10 00:30:30]

> I'm wondering about CertDatabase, it's platform specific and the overall
> InitSystemRequestContext slowness is specific to macs.

well less so since taking a look at what's in those .cc files