Investigate apparent data loss during flight

[jameysharp]

Once we turned on the GPS board, every few seconds the flight computer logged a group of "Sequence Error" (SEQE) messages. These show that we lost between 30ms and 1.75s of data each time (based on the size of the sequence number gaps), which is concerning.

We need to investigate:

1. Is there a corresponding gap of missing data prior to each SEQE message? (Probably, but it's worth checking.)
2. Does the gap span the same interval across all message types, or just the ones associated with sequence numbers, or just a few specific sources?
3. What caused these gaps?

[jameysharp]

After we turned on the GPS at timestamp 117244117823154, the first time this occurred was 47.059 seconds later, between timestamps 117291177275128 and 117291258831868, a span of 81.55674ms.

Here's that first sequence:

• last BMP1 at 117291172217781
• last RNHP at 117291176230024
• last JGPS at 117291177262417
• last ADIS at 117291177275128
• (gap)
• SEQE+first JGPS at 117291258831868: 162 packets dropped, ~80ms
• SEQE+first RNHP at 117291259391437: 82 packets dropped (not sure of nominal sample rate)
• SEQE+first ADIS at 117291259405475: 66 packets dropped, ~80.5ms
• SEQE+first BMP1 at 117291262243406: 8 packets dropped, ~80ms

Judging by these numbers, we can probably get pretty good estimates of gap times by looking at the number of JGPS packets dropped.

This consistency across sources is disappointing because it looks like we lost all data coming into the flight computer during these gaps. It isn't yet clear whether these were lost at the Ethernet switch, in the flight computer's NIC, or in the Linux kernel on the flight computer. But it does look like the fault was not in the microcontrollers, which appear to have been sending data reliably.