win: Add support for CPUTimes and StartTime to snapshot

win: Add support for CPUTimes and StartTime to snapshot

Follows https://codereview.chromium.org/1120383003/.

R=mark@chromium.org
BUG=crashpad:1
Committed: https://chromium.googlesource.com/crashpad/crashpad/+/06db728457693c3cce7e3fa8ecb63b675b256c04

[Mark Mentovai, 2015-05-04 21:29:24]

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_re...
File snapshot/win/process_reader_win.cc (right):

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_re...
snapshot/win/process_reader_win.cc:24: void TimerClear(timeval* tv) {
Do this in compat if it’s needed elsewhere, or just remove it if you don’t need
it in this file.

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_re...
File snapshot/win/process_reader_win.h (right):

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_re...
snapshot/win/process_reader_win.h:57: //!     failure, \a start_time will be set
to represent the epoch.
It’s OK to leave start_time undefined as far as the docs go on failure, even if
the implementation does in fact zero it. (But it doesn’t have to either.) Same
in CPUTimes().

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_sn...
File snapshot/win/process_snapshot_win.cc (right):

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_sn...
snapshot/win/process_snapshot_win.cc:102: process_reader_.StartTime(start_time);
INITIALIZATION_STATE_DCHECK_VALID(initialized_); here and in ProcessCPUTimes().

https://codereview.chromium.org/1119393003/diff/20001/util/win/time.cc
File util/win/time.cc (right):

https://codereview.chromium.org/1119393003/diff/20001/util/win/time.cc#newcode44
util/win/time.cc:44: timeval FiletimeToTimevalAmount(const FILETIME& filetime) {
I’m on the fence about having these two functions be totally distinct or share
more code. I guess it comes out in favor of it being OK for them to not share,
but if they grew any longer, things might change.

https://codereview.chromium.org/1119393003/diff/20001/util/win/time.h
File util/win/time.h (right):

https://codereview.chromium.org/1119393003/diff/20001/util/win/time.h#newcode23
util/win/time.h:23: //! \brief Convert Windows FILETIME to timeval, converting
from Windows epoch to
`FILETIME`, `timeval` in both comments.

[scottmg, 2015-05-05 16:45:48]

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_re...
File snapshot/win/process_reader_win.cc (right):

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_re...
snapshot/win/process_reader_win.cc:24: void TimerClear(timeval* tv) {
On 2015/05/04 21:29:23, Mark Mentovai wrote:
> Do this in compat if it’s needed elsewhere, or just remove it if you don’t
need
> it in this file.

Done.

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_re...
File snapshot/win/process_reader_win.h (right):

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_re...
snapshot/win/process_reader_win.h:57: //!     failure, \a start_time will be set
to represent the epoch.
On 2015/05/04 21:29:23, Mark Mentovai wrote:
> It’s OK to leave start_time undefined as far as the docs go on failure, even
if
> the implementation does in fact zero it. (But it doesn’t have to either.) Same
> in CPUTimes().

Done.

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_sn...
File snapshot/win/process_snapshot_win.cc (right):

https://codereview.chromium.org/1119393003/diff/20001/snapshot/win/process_sn...
snapshot/win/process_snapshot_win.cc:102: process_reader_.StartTime(start_time);
On 2015/05/04 21:29:24, Mark Mentovai wrote:
> INITIALIZATION_STATE_DCHECK_VALID(initialized_); here and in
ProcessCPUTimes().

Done.

https://codereview.chromium.org/1119393003/diff/20001/util/win/time.cc
File util/win/time.cc (right):

https://codereview.chromium.org/1119393003/diff/20001/util/win/time.cc#newcode44
util/win/time.cc:44: timeval FiletimeToTimevalAmount(const FILETIME& filetime) {
On 2015/05/04 21:29:24, Mark Mentovai wrote:
> I’m on the fence about having these two functions be totally distinct or share
> more code. I guess it comes out in favor of it being OK for them to not share,
> but if they grew any longer, things might change.

I was too. I made a MicrosecondsToTimeval helper.

https://codereview.chromium.org/1119393003/diff/20001/util/win/time.h
File util/win/time.h (right):

https://codereview.chromium.org/1119393003/diff/20001/util/win/time.h#newcode23
util/win/time.h:23: //! \brief Convert Windows FILETIME to timeval, converting
from Windows epoch to
On 2015/05/04 21:29:24, Mark Mentovai wrote:
> `FILETIME`, `timeval` in both comments.

Done.

[Mark Mentovai, 2015-05-05 19:53:37]

LGTM

https://codereview.chromium.org/1119393003/diff/40001/util/win/time.h
File util/win/time.h (right):

https://codereview.chromium.org/1119393003/diff/40001/util/win/time.h#newcode29
util/win/time.h:29: timeval FiletimeToTimevalAmount(const FILETIME& filetime);
Interval sounds snazzier than amount.

[scottmg, 2015-05-05 20:02:49]

Will land after the other CL that has the mini_chromium roll.

https://codereview.chromium.org/1119393003/diff/40001/util/win/time.h
File util/win/time.h (right):

https://codereview.chromium.org/1119393003/diff/40001/util/win/time.h#newcode29
util/win/time.h:29: timeval FiletimeToTimevalAmount(const FILETIME& filetime);
On 2015/05/05 19:53:36, Mark Mentovai wrote:
> Interval sounds snazzier than amount.

:) Done.

[scottmg, 2015-05-06 18:13:50]

Committed patchset #5 (id:80001) manually as
06db728457693c3cce7e3fa8ecb63b675b256c04 (presubmit successful).