Drop DecryptConfig::data_offset_.

Drop DecryptConfig::data_offset_.

This CL drops DecryptConfig::data_offset_ and moves the offset parsing code into WebM demuxers. This allows us to remove offset passing/calculation code in several layers of the media pipeline.

Background:
We used to have HMAC check in encrypted WebM RFC, which requires us to keep the IV together with the encrypted frame, hence the offset. Now the HMAC check is dropped from the RFC so offset is not necessary anymore.

BUG=298569
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=243672

[xhwang, 2013-12-30 23:46:56]

I'll update pepper in another CL. PTAL.

[xhwang, 2014-01-02 18:56:40]

+dalecurtis, PTAL!

[DaleCurtis, 2014-01-02 19:31:09]

Mechanically this lgtm % nits.  I defer to ddorwin@ for whether or not removing
the offset criteria is right for EME.

https://codereview.chromium.org/119203003/diff/130001/media/webm/webm_cluster...
File media/webm/webm_cluster_parser.cc (right):

https://codereview.chromium.org/119203003/diff/130001/media/webm/webm_cluster...
media/webm/webm_cluster_parser.cc:369: } else {
Does this not need a DecryptConfig?  Presumably we're not encrypted text.

https://codereview.chromium.org/119203003/diff/130001/media/webm/webm_crypto_...
File media/webm/webm_crypto_helpers.cc (right):

https://codereview.chromium.org/119203003/diff/130001/media/webm/webm_crypto_...
media/webm/webm_crypto_helpers.cc:29: scoped_ptr<DecryptConfig>* decrypt_config,
Just make it a regular pointer? This method shouldn't really care how the caller
allocates the DecryptConfig.

[DaleCurtis, 2014-01-02 19:39:34]

Oh, one thing though, provide some more details in the description about what
this change is doing (moving offset parsing into WebM demuxers versus being done
generically by the decryptor.

[ddorwin, 2014-01-07 23:56:27]

Removal LG, and I skimmed the CL. I defer to Dale on the details.

[xhwang, 2014-01-08 00:42:52]

On 2014/01/02 19:39:34, DaleCurtis wrote:
> Oh, one thing though, provide some more details in the description about what
> this change is doing (moving offset parsing into WebM demuxers versus being
done
> generically by the decryptor.

Done

[xhwang, 2014-01-08 00:43:35]

Comments only.

dalecurtis: Please take a look at my comments to see if you are happy with them
:)

https://codereview.chromium.org/119203003/diff/130001/media/webm/webm_cluster...
File media/webm/webm_cluster_parser.cc (right):

https://codereview.chromium.org/119203003/diff/130001/media/webm/webm_cluster...
media/webm/webm_cluster_parser.cc:369: } else {
On 2014/01/02 19:31:09, DaleCurtis wrote:
> Does this not need a DecryptConfig?  Presumably we're not encrypted text.

As you can see from l.314-l.319, encryption_key_id is only set for audio/video.

For now only audio/video can be encrypted. We may add support for text, but I am
not sure what's the use case for that :) Even if we need to do it, I'd keep it
in another CL.

https://codereview.chromium.org/119203003/diff/130001/media/webm/webm_crypto_...
File media/webm/webm_crypto_helpers.cc (right):

https://codereview.chromium.org/119203003/diff/130001/media/webm/webm_crypto_...
media/webm/webm_crypto_helpers.cc:29: scoped_ptr<DecryptConfig>* decrypt_config,
On 2014/01/02 19:31:09, DaleCurtis wrote:
> Just make it a regular pointer? This method shouldn't really care how the
caller
> allocates the DecryptConfig.

Hmm, we have "new" on l.53. If we return a regular pointer, the caller has to
remember to "delete" it, or wrap it in a scoped_ptr<>. I like to wrap it
immediately so that the ownership is super clear. Maybe I missed something?

[xhwang, 2014-01-08 00:44:58]

dmichael@: Please OWNERS review:
content/renderer/pepper/content_decryptor_delegate.cc

I'll clean up the pepper part in another CL.

[DaleCurtis, 2014-01-08 00:55:03]

still lgtm.

[dmichael (off chromium), 2014-01-08 17:51:25]

pepper lgtm

[commit-bot: I haz the power, 2014-01-08 17:53:09]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/xhwang@chromium.org/119203003/130001

[commit-bot: I haz the power, 2014-01-08 22:23:19]

Change committed as 243672