DRM: Fix potential race condition issues on setMediaKeys and quick content switching #1594
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
peaBerberian
added
bug
peaBerberian
force-pushed
the
fix/await-previous-setMediaKeys
branch
from
98d46ad
to
e66fea0
Compare
…ntent switching We've recently noticed on some LG TVs only that we triggered a `MEDIA_KEYS_ATTACHMENT_ERROR` if loading different contents very fast (e.g. "zapping" like crazy on the remote ~15-20 times on those). Turns out that the issue was due to a complex combination of optimizations and work-arounds, where we would "reset" the `MediaKeys` instance attached to the `HTMLMediaElement` (that's the work-around part of some LG issues) by calling the `setMediaKeys` API with a `null` `MediaKeys` (as documented in the EME recommendation), yet without awaiting that the original `setMediaKeys` call with the actual `MediaKeys` instance for that same content to be finished. Doing multiple `setMediaKeys` call without waiting for the previous one to finish leading to an error is an actually documented behavior in the EME recommendation so this could be categorized as a bug on our side. To fix this, I tried an easy way first but found out that there could be multiple other issues with `MediaKeys` attachment: if for whatever the still-pending `setMediaKeys` call for a previous content ends up failing when another content, which needs to rely on the same `MediaKeys` instance, is playing, no `MEDIA_KEYS_ATTACHMENT_ERROR` will be triggered and the RxPlayer could instead end up rebuffering indefinitely. So I decided to do a bigger code update: Now the method returning information on the `MediaKeys` attached to the `HTMLMediaElement` is always asynchronous (it returns a Promise) just in case the attachment task is still pending (and to also handle cases where it would fail). I still let a synchronous call for the `getKeySystemConfiguration` API, where it may thus also for now return the "expected" `MediaKeys` information (e.g. it will return the future expected state when attachment is still pending), because I wasn't sure of the behavior we want here - though I still am.
peaBerberian
force-pushed
the
fix/await-previous-setMediaKeys
branch
from
e66fea0
to
7311c34
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We've recently noticed on some LG TVs only that we triggered a
MEDIA_KEYS_ATTACHMENT_ERRORif loading different contents very fast (e.g. "zapping" like crazy on the remote ~15-20 times on those).Turns out that the issue was due to a complex combination of optimizations and work-arounds, where we would "reset" the
MediaKeysinstance attached to theHTMLMediaElement(that's the work-around part of some LG issues) by calling thesetMediaKeysAPI with anullMediaKeys(as documented in the EME recommendation), yet without awaiting that the originalsetMediaKeyscall with the actualMediaKeysinstance for that same content to be finished.Doing multiple
setMediaKeyscall without waiting for the previous one to finish leading to an error is an actually documented behavior in the EME recommendation so this could be categorized as a bug on our side.To fix this, I tried an easy way first but found out that there could be multiple other issues with
MediaKeysattachment: if for whatever reason the still-pendingsetMediaKeyscall for a previous content ends up failing when another content, which needs to rely on the sameMediaKeysinstance, is playing, noMEDIA_KEYS_ATTACHMENT_ERRORwill be triggered and the RxPlayer could instead end up rebuffering indefinitely.So I decided to do a bigger code update: Now the method returning information on the
MediaKeysattached to theHTMLMediaElementis always asynchronous (it returns a Promise) just in case the attachment task is still pending (and to also handle cases where it would fail).I still let a synchronous call for the
getKeySystemConfigurationAPI, where it may thus also for now return the "expected"MediaKeysinformation (e.g. it will return the future expected state when attachment is still pending), because I wasn't sure of the behavior we want here - though I still am.