runin-battery fails due to not enough discharge

[dsd]

Testing runin on 11.2.0 build 20 on XO-1.5.

The test always fails after some time because of runin-battery. fail.log says:

DISCHARGE
SOC diff: 18
FAIL: Battery: Net discharge of 18 is < 20

I know I could just manually tweak the file to evade this, but obviously Quanta wouldn't do that for every laptop. My question is, how is this supposed to work in the factory?

[rsmith]

Please attach the full set of logs make sure to include the pwr log so we can compare the power draw with runs on from 10.x. Either the power draw of the build running all the tests has changed or something is off with timing. Are you sure all the tests are running?

You can also prevent the battery test from stopping the runin procedure by setting BATTERY_NO_FAIL=TRUE

That will let the runin go through its cycle and we can see what the power draw is across the runin sequence.

[dsd]

same failure on 10.1.3

[dsd]

As a baseline I tested 10.1.3 on the same laptop. It failed in the same way. All logs attached as logs-SHC01601310-110530_053216.tar.gz

[dsd]

A debug setting had crept into the build (also for 10.1.3) meaning that it was suspending too much. fixed in olpc-runin-tests-0.9.44

[dsd]

test in 11.2.0 build 21

[dsd]

confirmed working on 11.2.0 build 22, it survived the whole 4 hours with no errors in logs.