Closed IvanLi1 closed 3 years ago
Haven't run into this situation before. So the device locks up when you try to read it through the root bus, but is ok if you read it through a mux bus? Shouldn't the EEPROM not be visible until you turn on the MUX?
This issue happens in the following situation while FruDevice is creating object path for EEPROM on DBus: e.g.
Interesting, I'd be fine with a change having the failure check be 1:1 with busnum, instead of root busnum. Obviously the assumption that there wouldn't be a re-use of a address on a root bus, that is on a mux bus, doesn't seem to work in your cases.
Is this still an issue?
Closing due to lack of response.
Commit : 7972bb95e4c9816217c025416701e48dae5b4f05 (Stop reading devices that don't like it)
Found that this commit will cause add-in card FRU behind MUX to be skipped when the EEPROM address on add-in card is the same as the address on root bus.
e.g. There's an address 0x80 on root bus(bus 0) and there's an EEPROM address 0x80(bus 31) on add-in card behind MUX. When failed to read bus 0 address 0x80 occurred, the address 0x80 on bus 31 won't be created on DBus and then cause user cannot read FRU content in add-on card via ipmitool fru print XX.
Could you help to comment on this issue.