LM5066I "POWER_GOOD" bit is likely defined incorrectly

[ericaasen]

While looking at Rev B Cosmos, we keep seeing that the fan HSCs are saying their PG is not set and I believe this happens on Rev A as well, though I don't have access to one that's powered up at the moment.

Even when the HSC seems to say that it's PG is not set by dumping the registers:

λ humility pmbus -r V54P5_FAN_CENTRAL -v
humility: attached via ST-Link V3

...

0x79 STATUS_WORD               0x0803
     |
     | b15    0b0 = no fault                 <= OutputVoltageFault
     | b13    0b0 = no fault                 <= InputFault
     | b12    0b0 = no fault                 <= FETFault
     | b11    0b1 = POWER_GOOD not set       <= PowerGoodStatus
     | b9     0b0 = no fault                 <= CircuitBreakerFault
     | b6     0b0 = power is not off         <= Off
     | b3     0b0 = no fault                 <= InputUndervoltageFault
     | b2     0b0 = no fault                 <= TemperatureFault
     | b1     0b1 = fault                    <= CMLFault
     | b0     0b1 = fault                    <= NoneOfTheAbove
     +-----------------------------------------------------------------------

when we measure pin 28 on the actual device, it is asserting PG as high (3.3V) on both sides of the isolator. I'm assuming this has something to do with this dumpster fire of a definition of POWER_GOOD in STATUS_WORD

[labbott]

I think this might be an issue in the PMBus crate itself https://github.com/oxidecomputer/pmbus/blob/master/src/lm5066i.ron#L121-L122