Occasionally mcp seems to not initialize the gyro readout properly. I've noticed it 2 times (out of more than 10 mcp restarts) in one night. All the individual gyro and combined gyro TOD are non-zero but flat. Other fast channels (EL, MC_EL_I_READ) appear to be read out as normal. Note in the example below that the derived field ERROR_AZ (AZ_GY-VEL_REQ_AZ) has no noise, which makes me think this is an mcp gyro communications problem rather than a defricher problem.
From the log mcp claims it has initialized the gyro communications:
0346-12-13 05:18:55.476 > Schedu: dsp1760.c:326 (initialize_dsp1760_interface):Initialized gyroscope interface
Then a few lines later:
= 0346-12-13 05:18:55.613 = Schedu: dsp1760.c:263 (dsp1760_process_data):Expected Sequence 1 but received 100 from gyro 1
I didn't see any other warnings or error messages from dsp1760.c.
Communication with the gyros can be recovered by restarting mcp.
Attempted to re-create. Observed in previous build. After commit cdcdb8c47a89e6d34bec468ed30830ee2b7556a9 the issue has not been observed in 50 restarts. If seen again, we'll re-open.
Occasionally mcp seems to not initialize the gyro readout properly. I've noticed it 2 times (out of more than 10 mcp restarts) in one night. All the individual gyro and combined gyro TOD are non-zero but flat. Other fast channels (EL, MC_EL_I_READ) appear to be read out as normal. Note in the example below that the derived field ERROR_AZ (AZ_GY-VEL_REQ_AZ) has no noise, which makes me think this is an mcp gyro communications problem rather than a defricher problem.
From the log mcp claims it has initialized the gyro communications:
I didn't see any other warnings or error messages from dsp1760.c.
Communication with the gyros can be recovered by restarting mcp.