Collect total counters in case of GDI failure.

Collect total counters in case of GDI failure.

If a system instability is caused by some non-browser process, it will
be known from the total usage counters when crash dump is received.

Thus we can distinguish a case when some non-browser process is
misbehaving from other kinds of system resource exhaustion.

BUG=661461
R=thestig@chromium.org

Committed: https://crrev.com/9625fddb94caf89f7f967326bb662aced7b2d884
Cr-Commit-Position: refs/heads/master@{#430161}

[anpol, 2016-11-02 13:30:45]

Description was changed from

==========
Collect total counters in case of GDI failure.

If a system instability is caused by some non-browser process, it will
be known from the total usage counters when crash dump is received.

Thus we can distinguish a case when some non-browser process is
misbehaving from other kinds of system resource exhaustion.

BUG=661461
R=thestig@chromium.org
==========

to

==========
Collect total counters in case of GDI failure.

If a system instability is caused by some non-browser process, it will
be known from the total usage counters when crash dump is received.

Thus we can distinguish a case when some non-browser process is
misbehaving from other kinds of system resource exhaustion.

BUG=661461
R=thestig@chromium.org
==========

[anpol, 2016-11-02 13:30:45]

anpol@yandex-team.ru changed reviewers:
+ ananta@chromium.org, cpu@chromium.org, pkasting@chromium.org,
siggi@chromium.org

[anpol, 2016-11-03 12:07:00]

PTAL

[Sigurður Ásgeirsson, 2016-11-03 17:30:20]

This looks very sane to me - there's some additional cost, but since we're about
to die, I don't think that's a concern :).

lgtm

[Lei Zhang, 2016-11-03 18:53:12]

lgtm

https://codereview.chromium.org/2467283002/diff/1/base/debug/gdi_debug_util_w...
File base/debug/gdi_debug_util_win.cc (right):

https://codereview.chromium.org/2467283002/diff/1/base/debug/gdi_debug_util_w...
base/debug/gdi_debug_util_win.cc:60: if (total_peak_gdi_count < num_gdi_handles)
total_peak_gdi_count = std::max(total_peak_gdi_count, num_gdi_handles);

Also line 70, if you don't mind.

[Peter Kasting, 2016-11-03 21:35:40]

Going to take myself off this since you have other review coverage.

For future reference, when adding multiple reviewers:
* Always say what you want each to review, and in what role (e.g. reviewer vs.
OWNER)
* Generally minimize the number of overlapping reviewers, e.g. just get an
OWNER's review unless another person would be more familiar with the change in
question in a way that's important

[Peter Kasting, 2016-11-03 21:35:49]

pkasting@chromium.org changed reviewers:
- pkasting@chromium.org

[anpol, 2016-11-04 07:24:07]

On 2016/11/03 21:35:40, Peter Kasting wrote:
> Going to take myself off this since you have other review coverage.
> 
> For future reference, when adding multiple reviewers:
> * Always say what you want each to review, and in what role (e.g. reviewer vs.
> OWNER)
> * Generally minimize the number of overlapping reviewers, e.g. just get an
> OWNER's review unless another person would be more familiar with the change in
> question in a way that's important

OK, thank you.

It would be great if cpu takes a look at these changes, as he is the author of
the original code.

[anpol, 2016-11-05 08:55:43]

The CQ bit was checked by anpol@yandex-team.ru

[anpol, 2016-11-05 08:55:43]

The patchset sent to the CQ was uploaded after l-g-t-m from
thestig@chromium.org, siggi@chromium.org
Link to the patchset: https://codereview.chromium.org/2467283002/#ps20001
(title: "Use std::max instead of 'if'.")

[commit-bot: I haz the power, 2016-11-05 08:55:53]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-05 09:56:47]

Description was changed from

==========
Collect total counters in case of GDI failure.

If a system instability is caused by some non-browser process, it will
be known from the total usage counters when crash dump is received.

Thus we can distinguish a case when some non-browser process is
misbehaving from other kinds of system resource exhaustion.

BUG=661461
R=thestig@chromium.org
==========

to

==========
Collect total counters in case of GDI failure.

If a system instability is caused by some non-browser process, it will
be known from the total usage counters when crash dump is received.

Thus we can distinguish a case when some non-browser process is
misbehaving from other kinds of system resource exhaustion.

BUG=661461
R=thestig@chromium.org
==========

[commit-bot: I haz the power, 2016-11-05 09:56:47]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-11-05 09:59:15]

Description was changed from

==========
Collect total counters in case of GDI failure.

If a system instability is caused by some non-browser process, it will
be known from the total usage counters when crash dump is received.

Thus we can distinguish a case when some non-browser process is
misbehaving from other kinds of system resource exhaustion.

BUG=661461
R=thestig@chromium.org
==========

to

==========
Collect total counters in case of GDI failure.

If a system instability is caused by some non-browser process, it will
be known from the total usage counters when crash dump is received.

Thus we can distinguish a case when some non-browser process is
misbehaving from other kinds of system resource exhaustion.

BUG=661461
R=thestig@chromium.org

Committed: https://crrev.com/9625fddb94caf89f7f967326bb662aced7b2d884
Cr-Commit-Position: refs/heads/master@{#430161}
==========

[commit-bot: I haz the power, 2016-11-05 09:59:16]

Patchset 2 (id:??) landed as
https://crrev.com/9625fddb94caf89f7f967326bb662aced7b2d884
Cr-Commit-Position: refs/heads/master@{#430161}