Bugzilla – Bug 9058
Ocassionall panic due to incorrect AC line voltage measurement
Last modified: 2009-09-08 09:19:09 UTC
In the latest firmware there is a bug which causes the firmware to sometimes incorrectly read the line voltage monitoring ADC. Probably a race condition with the other SPI config interface, which share one of those pins.
Sean: are you planning on fixing this in 7.2?
working on it now
I can't figure out why we get bad readings. It's pretty clear that it's something to do with the changes around V. 43 or so, but the spinlock code looks correct. I added a sanity check to require two consistent consecutive readings, which should prevent the over-voltage panic. I can't say for sure if this bug is actually fixed because it could take thousands of hours to reproduce. We'll have to keep an eye on this, and we can close it after it's been proven in the field for a month. Leaving the bug open as a reminder.
Sean: Do you consider this fixed with your change?
Yes but I can't test it. We should still track this issue in some way...
Chris: How do you want to track this? We should do some basic verification of the voltage, then keep an eye on it post 7.2.
Fixed in firmware 59.
I'll check and see what we have for transporter fw test cases and throw something in there.