Add support for v0.3 of the encrypted WebM specification. (Revert Fix)

Add support for v0.3 of the encrypted WebM specification. (Revert Fix)

- Fixed HmacCheckFailure from aes_decryptor_unittest.cc that was causing a
failure on the ASAN bots.
- Added code to handle the signal_byte contained within WebM
encrypted Blocks.
- Added a unittest to aes_decryptor to handle an encrypted WebM
Block with an unencrypted frame.
- Added two unittests to aes_decryptor to try an encrypted WebM
frame as an unencrypted frame and an unencrypted WebM frame as an
encrypted frame.

Revert "Revert 149449 - Add support for v0.3 of the encrypted WebM specification."

This reverts commit 672f667ed68e59ec0aaf5c40167e890673ef3f7c.

BUG=139876
TEST=Run media_unittests --gtest_filter=AesDecryptor* and all
tests must pass. media_unittests must pass on linux_asan and mac_asan.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=149696

[xhwang, 2012-08-01 23:15:39]

LGTM w/ nits.

http://codereview.chromium.org/10824136/diff/1/media/crypto/aes_decryptor_uni...
File media/crypto/aes_decryptor_unittest.cc (right):

http://codereview.chromium.org/10824136/diff/1/media/crypto/aes_decryptor_uni...
media/crypto/aes_decryptor_unittest.cc:339: wrong_key[0]++;
nit: I feel we have too many of this scoped_array/memcpy pattern in
AesDecryptorTest. But I don't find a better solution either. If you have a
better idea, improve it, or just leave it as is.

http://codereview.chromium.org/10824136/diff/1/media/crypto/aes_decryptor_uni...
media/crypto/aes_decryptor_unittest.cc:484: scoped_array<uint8> frame_change(new
uint8[frame.encrypted_data_size]);
nit: Can this name be more specific? e.g. frame_with_wrong_signal_byte or
signal_byte_altered_frame or something better?

http://codereview.chromium.org/10824136/diff/1/media/crypto/aes_decryptor_uni...
media/crypto/aes_decryptor_unittest.cc:503: scoped_array<uint8> frame_change(new
uint8[frame.encrypted_data_size]);
ditto

[fgalligan1, 2012-08-01 23:49:58]

https://chromiumcodereview.appspot.com/10824136/diff/1/media/crypto/aes_decry...
File media/crypto/aes_decryptor_unittest.cc (right):

https://chromiumcodereview.appspot.com/10824136/diff/1/media/crypto/aes_decry...
media/crypto/aes_decryptor_unittest.cc:339: wrong_key[0]++;
On 2012/08/01 23:15:39, xhwang wrote:
> nit: I feel we have too many of this scoped_array/memcpy pattern in
> AesDecryptorTest. But I don't find a better solution either. If you have a
> better idea, improve it, or just leave it as is.

Changed to vector<uint8>, which removes the memcpy line.

https://chromiumcodereview.appspot.com/10824136/diff/1/media/crypto/aes_decry...
media/crypto/aes_decryptor_unittest.cc:484: scoped_array<uint8> frame_change(new
uint8[frame.encrypted_data_size]);
On 2012/08/01 23:15:39, xhwang wrote:
> nit: Can this name be more specific? e.g. frame_with_wrong_signal_byte or
> signal_byte_altered_frame or something better? 

Done.

https://chromiumcodereview.appspot.com/10824136/diff/1/media/crypto/aes_decry...
media/crypto/aes_decryptor_unittest.cc:503: scoped_array<uint8> frame_change(new
uint8[frame.encrypted_data_size]);
On 2012/08/01 23:15:39, xhwang wrote:
> ditto

Done.

[xhwang, 2012-08-02 00:00:22]

http://codereview.chromium.org/10824136/diff/1/media/crypto/aes_decryptor_uni...
File media/crypto/aes_decryptor_unittest.cc (right):

http://codereview.chromium.org/10824136/diff/1/media/crypto/aes_decryptor_uni...
media/crypto/aes_decryptor_unittest.cc:339: wrong_key[0]++;
On 2012/08/01 23:49:58, fgalligan1 wrote:
> On 2012/08/01 23:15:39, xhwang wrote:
> > nit: I feel we have too many of this scoped_array/memcpy pattern in
> > AesDecryptorTest. But I don't find a better solution either. If you have a
> > better idea, improve it, or just leave it as is.
> 
> Changed to vector<uint8>, which removes the memcpy line.

Why wrong_key is not vector<uint8> ?

[fgalligan1, 2012-08-02 00:13:30]

https://chromiumcodereview.appspot.com/10824136/diff/1/media/crypto/aes_decry...
File media/crypto/aes_decryptor_unittest.cc (right):

https://chromiumcodereview.appspot.com/10824136/diff/1/media/crypto/aes_decry...
media/crypto/aes_decryptor_unittest.cc:339: wrong_key[0]++;
On 2012/08/02 00:00:22, xhwang wrote:
> On 2012/08/01 23:49:58, fgalligan1 wrote:
> > On 2012/08/01 23:15:39, xhwang wrote:
> > > nit: I feel we have too many of this scoped_array/memcpy pattern in
> > > AesDecryptorTest. But I don't find a better solution either. If you have a
> > > better idea, improve it, or just leave it as is.
> > 
> > Changed to vector<uint8>, which removes the memcpy line.
> 
> Why wrong_key is not vector<uint8> ?

Missed it before I left. Fixed.

[xhwang, 2012-08-02 00:15:01]

lgtm

[commit-bot: I haz the power, 2012-08-02 19:24:39]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/fgalligan@chromium.org/10824136/6011

[commit-bot: I haz the power, 2012-08-02 20:57:41]

Change committed as 149696