barawn
commented
3 years ago nope, it definitely would do something. The FIFO read latency is only a single clock, so instead of capturing the first entry in the FIFO it'd move on to the second one if it's filled.
The event header read signals weren't qualified on ack_o being asserted, so the read enable signals were high for two clocks.
This tends to not be a problem because there are almost always zero pending events in the header, so the double-read does nothing. But if a new trigger occurs between the trigger busy being cleared and the headers being read out (~few 100 ns) the FIFO advances twice.
Still don't think this does anything, but it might do something in conjuction with other issues.