Fix Windows HID buffer errors.

Fix Windows HID buffer errors.

Several incorrect assumptions about actual and expected buffer
sizes are corrected in the Windows HidConnection implementation.

BUG=369344
TBR=rpaquay

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=268215

[Ken Rockot(use gerrit already), 2014-05-02 00:16:55]

Could you please take a look at this too? I'm not sure there's much anyone can
actually review without manually testing to verify the changes.

[Ken Rockot(use gerrit already), 2014-05-02 00:18:19]

https://codereview.chromium.org/261053005/diff/1/device/hid/hid_connection_wi...
File device/hid/hid_connection_win.cc (right):

https://codereview.chromium.org/261053005/diff/1/device/hid/hid_connection_wi...
device/hid/hid_connection_win.cc:190: new PendingHidTransfer(this,
output_buffer, NULL, callback));
Apart from the buffer size corrections, this was just flat out wrong and would
frequently result in a use-after free that Windows apparently doesn't segfault
on (instead it just fails on the IO call). Passing output_buffer here (instead
of buffer) will always correctly retain a ref to the buffer.

[rpaquay, 2014-05-05 15:41:08]

lgtm

[Ken Rockot(use gerrit already), 2014-05-05 16:02:24]

The CQ bit was checked by rockot@chromium.org

[commit-bot: I haz the power, 2014-05-05 16:02:59]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/rockot@chromium.org/261053005/1

[commit-bot: I haz the power, 2014-05-05 17:48:03]

Change committed as 268215

[jracle (use Gerrit), 2014-05-06 15:02:54]

https://codereview.chromium.org/261053005/diff/1/device/hid/hid_connection_wi...
File device/hid/hid_connection_win.cc (right):

https://codereview.chromium.org/261053005/diff/1/device/hid/hid_connection_wi...
device/hid/hid_connection_win.cc:190: new PendingHidTransfer(this,
output_buffer, NULL, callback));
Hi Ken, thanks very much for fixing that one!
I also found it while debugging IO failure..