ppapi: VideoEncoder: prevent scheduling encoding multiple times

ppapi: VideoEncoder: prevent scheduling encoding multiple times

The ScheduleNextEncode() can be called at multiple stages in the
example depending on whether it needs to reconfigure the
MediaVideoStreamTrack. This can lead the example to encode at twice
30fps. At this speed we can exhaust the 33ms lifespan of video
buffers, leading to errors in the example plugins which can't retrieve
new frames to respect the 30fps framerate.

BUG=503153
TEST=run video_encoder example from NaCl SDK with VP8 codec and verify there are no error messages after 1 minute

Committed: https://crrev.com/cd74d5568a25c0a50ca8275b88a8d37bc066d275
Cr-Commit-Position: refs/heads/master@{#336850}

[llandwerlin-old, 2015-06-30 14:06:23]

lionel.g.landwerlin@intel.com changed reviewers:
+ bbudge@chromium.org, sbc@google.com

[llandwerlin-old, 2015-06-30 14:06:23]

bbudge@: Could you review this change in ppapi/ ? Thanks.

sbc@: Just need a stamp after bbudge's review, this is the same code copied into
the NaCl SDK. Feel free to do proper review if you want. Thanks.

[llandwerlin-old, 2015-06-30 14:08:01]

lionel.g.landwerlin@intel.com changed reviewers:
+ sbc@chromium.org
- sbc@google.com

[bbudge, 2015-06-30 15:45:05]

lgtm

[binji, 2015-06-30 17:02:02]

binji@chromium.org changed reviewers:
+ binji@chromium.org

[binji, 2015-06-30 17:02:02]

https://codereview.chromium.org/1218513003/diff/1/native_client_sdk/src/examp...
File native_client_sdk/src/examples/api/video_encode/video_encode.cc (right):

https://codereview.chromium.org/1218513003/diff/1/native_client_sdk/src/examp...
native_client_sdk/src/examples/api/video_encode/video_encode.cc:355: PP_Time
delta =
This produces a constant value for delta:

1. The inner clamp generates a value in the range [0, tick].
2. Then you subtract tick, so the value is in the range [-tick, 0].
3. Then you clamp, so the value is in the range [0, 0].
4. Then you subtract this from tick, so the value of delta is always tick.

[llandwerlin-old, 2015-06-30 17:15:23]

https://codereview.chromium.org/1218513003/diff/1/native_client_sdk/src/examp...
File native_client_sdk/src/examples/api/video_encode/video_encode.cc (right):

https://codereview.chromium.org/1218513003/diff/1/native_client_sdk/src/examp...
native_client_sdk/src/examples/api/video_encode/video_encode.cc:355: PP_Time
delta =
On 2015/06/30 17:02:02, binji wrote:
> This produces a constant value for delta:
> 
> 1. The inner clamp generates a value in the range [0, tick].
> 2. Then you subtract tick, so the value is in the range [-tick, 0].
> 3. Then you clamp, so the value is in the range [0, 0].
> 4. Then you subtract this from tick, so the value of delta is always tick.

Thanks, just clamping once between [0, tick].

[binji, 2015-06-30 18:50:34]

lgtm

https://codereview.chromium.org/1218513003/diff/20001/native_client_sdk/src/e...
File native_client_sdk/src/examples/api/video_encode/video_encode.cc (right):

https://codereview.chromium.org/1218513003/diff/20001/native_client_sdk/src/e...
native_client_sdk/src/examples/api/video_encode/video_encode.cc:357: PP_Time
delta = tick - clamp(0, tick, now - last_encode_tick_ - tick);
Just to be clear: this assumes the callback is never called early, right? I
guess that never happens in practice.

[llandwerlin-old, 2015-06-30 19:14:15]

https://codereview.chromium.org/1218513003/diff/20001/native_client_sdk/src/e...
File native_client_sdk/src/examples/api/video_encode/video_encode.cc (right):

https://codereview.chromium.org/1218513003/diff/20001/native_client_sdk/src/e...
native_client_sdk/src/examples/api/video_encode/video_encode.cc:357: PP_Time
delta = tick - clamp(0, tick, now - last_encode_tick_ - tick);
On 2015/06/30 18:50:34, binji wrote:
> Just to be clear: this assumes the callback is never called early, right? I
> guess that never happens in practice.

Yes, that's right. I assumed the implementation would be similar to a poll() on
unix ("the actual timeout interval shall be rounded up to the next supported
value" http://pubs.opengroup.org/onlinepubs/009695399/functions/poll.html)

[llandwerlin-old, 2015-06-30 19:14:30]

The CQ bit was checked by lionel.g.landwerlin@intel.com

[llandwerlin-old, 2015-06-30 19:14:31]

The patchset sent to the CQ was uploaded after l-g-t-m from bbudge@chromium.org
Link to the patchset: https://codereview.chromium.org/1218513003/#ps20001
(title: "Do not clamp twice")

[commit-bot: I haz the power, 2015-06-30 19:14:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1218513003/20001

[commit-bot: I haz the power, 2015-06-30 19:14:54]

The author lionel.g.landwerlin@intel.com has not signed Google Contributor
License Agreement. Please visit https://cla.developers.google.com to sign and
manage CLA.

[commit-bot: I haz the power, 2015-06-30 19:24:33]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-06-30 19:26:41]

Patchset 2 (id:??) landed as
https://crrev.com/cd74d5568a25c0a50ca8275b88a8d37bc066d275
Cr-Commit-Position: refs/heads/master@{#336850}