Try NtDuplicateObject instead of DuplicateHandle on Windows 8/8.1 for debugging AppContainer relate…

Try NtDuplicateObject instead of DuplicateHandle on Windows 8/8.1 for debugging AppContainer related failures.
This patch should be reverted as soon as we get some confirmation that NtDuplicateObject is working better compared to DuplicateHandle.
Theory behind this patch is that it might be possible that on machines which reporting this find of failure have some software like AV which might be intercepting calls to DuplciateHandle and may be failing somewhere in intercepted code path due to AppContainer.
Looking in IDA error code that we are receiving 0xC0000023 (Which translates INSUFFICIENT_BUFFER (0x7a) in win32 language) doesn't seem to be in NtDuplicateObject code path (At least not found easily.).

BUG=468922
R=jschuh@chromium.org,cpu@chromium.org

Committed: https://crrev.com/f7d692ca5448140fb21ffa42ca8b1535aae0b490
Cr-Commit-Position: refs/heads/master@{#326943}

[Shrikant Kelkar, 2015-04-23 21:40:25]

ptal.

[cpu_(ooo_6.6-7.5), 2015-04-24 02:39:56]

https://codereview.chromium.org/1101913002/diff/1/base/win/scoped_process_inf...
File base/win/scoped_process_information.cc (right):

https://codereview.chromium.org/1101913002/diff/1/base/win/scoped_process_inf...
base/win/scoped_process_information.cc:31: typedef LONG NTSTATUS;
just do LONG w/o typedef

https://codereview.chromium.org/1101913002/diff/1/base/win/scoped_process_inf...
base/win/scoped_process_information.cc:52: RtlNtStatusToDosError
ntstatus_to_doserror =
do we need this? I don't where?

[Shrikant Kelkar, 2015-04-24 18:31:27]

https://codereview.chromium.org/1101913002/diff/1/base/win/scoped_process_inf...
File base/win/scoped_process_information.cc (right):

https://codereview.chromium.org/1101913002/diff/1/base/win/scoped_process_inf...
base/win/scoped_process_information.cc:31: typedef LONG NTSTATUS;
On 2015/04/24 02:39:56, cpu wrote:
> just do LONG w/o typedef

Done.

https://codereview.chromium.org/1101913002/diff/1/base/win/scoped_process_inf...
base/win/scoped_process_information.cc:52: RtlNtStatusToDosError
ntstatus_to_doserror =
On 2015/04/24 02:39:56, cpu wrote:
> do we need this? I don't where?

kernelbase::DuplicateHandle calls NtDupliateObject and sets up last error after
receiving NTSTATUS. Tried to keep the same behavior. One more reason is that in
caller chain we report histogram value which indicates GetLastError(), so during
this experiment we can find out if that value changes as well.

[cpu_(ooo_6.6-7.5), 2015-04-24 21:58:27]

lgtm

[Shrikant Kelkar, 2015-04-24 21:59:30]

The CQ bit was checked by shrikant@chromium.org

[commit-bot: I haz the power, 2015-04-24 22:00:12]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1101913002/20001

[commit-bot: I haz the power, 2015-04-25 00:36:01]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-04-25 00:36:54]

Patchset 2 (id:??) landed as
https://crrev.com/f7d692ca5448140fb21ffa42ca8b1535aae0b490
Cr-Commit-Position: refs/heads/master@{#326943}