Zeroes buffer for DocumentProperti

Zeroes buffer for DocumentProperties.
Suspecting that driver does not reset dmDriverExtra.

BUG=421402
Committed: https://crrev.com/f16d34bbc32d79d739bd2e151552e5229de5e3c1
Cr-Commit-Position: refs/heads/master@{#298760}

[Vitaly Buka (NO REVIEWS), 2014-10-08 20:03:25]

vitalybuka@chromium.org changed reviewers:
+ thestig@chromium.org

[Lei Zhang, 2014-10-08 20:11:34]

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper.cc
File printing/backend/win_helper.cc (right):

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper....
printing/backend/win_helper.cc:481: memset(out.get(), 0, buffer_size);
Just calloc() instead?

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper....
printing/backend/win_helper.cc:490: CHECK_GE(buffer_size, size + extra_size);
Do we care if size + extra_size overflows?

[Vitaly Buka (NO REVIEWS), 2014-10-08 20:18:15]

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper.cc
File printing/backend/win_helper.cc (right):

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper....
printing/backend/win_helper.cc:481: memset(out.get(), 0, buffer_size);
On 2014/10/08 20:11:34, Lei Zhang wrote:
> Just calloc() instead?

Done.

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper....
printing/backend/win_helper.cc:490: CHECK_GE(buffer_size, size + extra_size);
On 2014/10/08 20:11:34, Lei Zhang wrote:
> Do we care if size + extra_size overflows?

Done.

[Lei Zhang, 2014-10-08 20:24:16]

lgtm

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper.cc
File printing/backend/win_helper.cc (right):

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper....
printing/backend/win_helper.cc:490: CHECK_GE(buffer_size, size + extra_size);
On 2014/10/08 20:18:15, Vitaly Buka wrote:
> On 2014/10/08 20:11:34, Lei Zhang wrote:
> > Do we care if size + extra_size overflows?
> 
> Done.

I just checked and a WORD is only 16-bit. There's no way 2 WORDS added together
can overflow a LONG. So why not just: LONG actual_size = out->dmSize +
out->dmDriverExtra; ?

[Vitaly Buka (NO REVIEWS), 2014-10-08 20:24:28]

The CQ bit was checked by vitalybuka@chromium.org

[Vitaly Buka (NO REVIEWS), 2014-10-08 20:24:32]

The CQ bit was unchecked by vitalybuka@chromium.org

[Vitaly Buka (NO REVIEWS), 2014-10-08 20:32:56]

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper.cc
File printing/backend/win_helper.cc (right):

https://codereview.chromium.org/644463002/diff/1/printing/backend/win_helper....
printing/backend/win_helper.cc:490: CHECK_GE(buffer_size, size + extra_size);
C++ does WORD+WORD->WORD, so old version indeed had overflow problem.
However it made no difference because overflow would likely satisfy CHECK_GE

LONG would work instead of int, but we don't need LONG for api calls like above,
so int should be used, according our style.

Also I put two local variable hoping to see them in future crash dumps on stack.


On 2014/10/08 20:24:15, Lei Zhang wrote:
> On 2014/10/08 20:18:15, Vitaly Buka wrote:
> > On 2014/10/08 20:11:34, Lei Zhang wrote:
> > > Do we care if size + extra_size overflows?
> > 
> > Done.
> 
> I just checked and a WORD is only 16-bit. There's no way 2 WORDS added
together
> can overflow a LONG. So why not just: LONG actual_size = out->dmSize +
> out->dmDriverExtra; ?

[Vitaly Buka (NO REVIEWS), 2014-10-08 22:07:56]

The CQ bit was checked by vitalybuka@chromium.org

[commit-bot: I haz the power, 2014-10-08 22:11:44]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/644463002/20001

[Vitaly Buka (NO REVIEWS), 2014-10-09 00:17:15]

The CQ bit was unchecked by vitalybuka@chromium.org

[Vitaly Buka (NO REVIEWS), 2014-10-09 00:17:25]

The CQ bit was checked by vitalybuka@chromium.org

[commit-bot: I haz the power, 2014-10-09 00:21:56]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/644463002/20001

[commit-bot: I haz the power, 2014-10-09 04:33:06]

Committed patchset #2 (id:20001) as 4372c0acb01d253c3c3b931e87da1e4c20018521

[commit-bot: I haz the power, 2014-10-09 04:33:58]

Patchset 2 (id:??) landed as
https://crrev.com/f16d34bbc32d79d739bd2e151552e5229de5e3c1
Cr-Commit-Position: refs/heads/master@{#298760}