* This patch adds a new microsecond-resolution function call, LiGetMicroseconds(), to complement
the existing LiGetMillis(). Many variables used by stats have been updated to work at this
higher resolution and now provide better results when displaying e.g. sub-millisecond frametime stats.
To try and avoid confusion, variables that now contain microseconds have been renamed with a suffix
of 'Us', and those ending in 'Ms' contain milliseconds. I originally experimented with nanoseconds but it
felt like overkill for our needs.
Public API in Limelight.h:
uint64_t LiGetMicroseconds(void);
uint64_t LiGetMillis(void);
const RTP_AUDIO_STATS* LiGetRTPAudioStats(void); // provides access to RTP data for the overlay stats
const RTP_VIDEO_STATS* LiGetRTPVideoStats(void);
Note: Users of this library may need to make changes. If using LiGetMillis() to track the duration of
something that is shown to the user, consider switching to LiGetMicroseconds(). Remember to divide by
1000 at time of display to show in milliseconds.
It's possible for a sequence of packets seqnums like 2, 0, 1, 2 to end up hitting the fast path
again after filling in gap of OOS packets. The fast path assumes duplicate packets are caught
by the next contiguous sequence number check, but this is only true if that value is kept updated.
Since the slow path doesn't update the next contiguous sequence number, it's no longer safe to
use the fast path for the remaining packets in the frame because duplicate packets would be
mishandled.
Queuing duplicate packets violates all sorts of queue assumptions and can cause us to return
corrupt data to the decoder or attempt an FEC recovery without enough shards.
Since we always allocate fixed size these aren't likely to be exploitable,
but we ought to fix them anyway. Worst case, we will just read some garbage
and generate corrupt video output.
With predictive RFI, we can recover from a lost frame as soon as the next frame
by predicting whether we will have enough FEC data to be be able to recover a
frame based on the sequence numbers of the received packets.
Cameron Gutman
Andy Grundman
ns6089