Forked from  bug #671374 

Someone saw this on a minnie:

<4>[22450.787789] cros-ec-spi spi0.0: EC failed to respond in time
<3>[22450.788035] cros-ec-spi spi0.0: Command xfer error (err:-110)
<3>[22450.788048] bq27x00-battery 20-0055: error reading voltage
<3>[22450.788061] power_supply bq27500-0: driver failed to report `voltage_now' property: -110

* AP/EC comms are supposed to be robust, so we shouldn't be getting transfer errors.

* Possibly this isn't actually a failure in AP/EC comms (hard to tell without seeing the EC logs) but was just the battery doing a crazy long clock stretch in i2c passthrough?

* Plausibly we could add retries to the bq27500 driver (like SBS battery has), but I don't terribly like it.

---

No way to reproduce as far as I know.  We could:

* Try to reproduce

* Attempt some type of speculative workaround.

* Add some stats to see how often we're seeing this in reality, which might help us tell if the speculative workaround worked.

---

I won't have time to drive this in the short term.  Sameer / Stefan: any candidates for someone who might want this?