fenugrec
commented
1 year ago wow good find re the wrong inner/outer index var ! That is a very dusty part of the code...
I'm surprised we don't already have a test for this, was expecting Ctest to start failing after this patch. Would you be able to add coverage for this ?
aaeegg
This fixes a few bugs in the "test readiness" command and adds a summary count to its output.
The "test readiness" command was displaying the readiness monitor status backwards: complete monitors were displayed as not complete, and vice versa. This was due to cmd_test_readiness interpreting a set bit in the Mode 1 PID $01 response to mean that the monitor is complete, but actually, a set bit means that the monitor is incomplete. (I mentioned this on IRC a few years ago, and I think fenugrec and I each assumed the other was going to fix it, so here we are...)
My freshly reset bench ECU responds to Mode 1 PID $01 with 41 01 00 07 ED ED. That decodes as follows, per SAE J1979 APR2002 Tables 17 and B2:
freediag was reporting the readiness status wrong:
With the change, it gets it right:
But note the weird ECU 0, 1, 2, 4 prefix. What's going on there is that, at least in theory, multiple ECUs can respond to the same Mode 1 PID $01 request, and freediag tries to iterate over all the replies it got and print all of them. But the inner loop, which iterates over the bits for each monitor, uses the same index variable as the outer loop, which iterates over the responses! So it prints the monitor index as the ECU number, and also, the outer loop quits after the first iteration so if there actually were multiple responses, it only prints the first. Adding a second index variable fixes this:
I also added a summary display at the end showing how many continuous and non-continuous monitors were incomplete and the total number of incomplete monitors:
This allows for rapid assessment of whether a car is ready for emissions testing. The exact number of incomplete monitors allowed varies by jurisdiction and model year.
Note that in do_j1979_getdtcs, the initial "scan" command also looks at the Mode 1 PID $01 response and prints a message if there are incomplete monitors. But the existing code there is already correct.