Hi, I recently built a machine with a Threadripper 3970X and a Gigabyte TRX40 Aorus pro wifi, and while it works for most of everything with 5.5 rc3 or rc4, I do have an issue with SMBus support, which I was supposing is why I'm not seeing any readings for e.g. CPU fans with lm-sensors. The i2c-piix4 driver loads, but does nothing, leaving this message in dmesg: [ 2.711213] ACPI Warning: SystemIO range 0x0000000000000B00-0x0000000000000B08 conflicts with OpRegion 0x0000000000000B00-0x0000000000000B0F (\GSA1.SMBI) (20191018/utaddress-213) [ 2.711218] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver Instrumenting the module, I can see it comes from https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/i2c/busses/i2c-piix4.c?id=fd6988496e79a6a4bdb514a4655d2920209eb85d#n323 If I remove the check, the module loads and dmesg says: [ 7589.210039] piix4_smbus 0000:00:14.0: SMBus Host Controller at 0xb00, revision 0 [ 7589.210041] piix4_smbus 0000:00:14.0: Using register 0x02 for SMBus port selection and now i2cdetect -l does find the smbus adapter, which it didn't find before the change. i2c-6 smbus SMBus PIIX4 adapter port 2 at 0b00 SMBus adapter i2c-5 smbus SMBus PIIX4 adapter port 0 at 0b00 SMBus adapter sensors-detect doesn't find any sensors in there, though (only SPD EEPROMs), so that may not be where my missing sensors are. In any case, the SMBus adapter is currently not initialized without tweaking the module. I should add, though, that during boot, I see a bunch of these in dmesg: [ 0.388399] ACPI BIOS Error (bug): Failure creating named object [\_SB.I2CA.WT1A], AE_ALREADY_EXISTS (20191018/dswload2-324) [ 0.388405] ACPI Error: AE_ALREADY_EXISTS, During name lookup/catalog (20191018/psobject-221) [ 0.388407] ACPI: Skipping parse of AML opcode: OpcodeName unavailable (0x5B82) [ 0.388408] ACPI BIOS Error (bug): Failure creating named object [\_SB.I2CA.MT1A], AE_ALREADY_EXISTS (20191018/dswload2-324) [ 0.388411] ACPI Error: AE_ALREADY_EXISTS, During name lookup/catalog (20191018/psobject-221) [ 0.388412] ACPI: Skipping parse of AML opcode: OpcodeName unavailable (0x5B82) (and many more similar ones) I don't know if they'd be related, but dumping the ACPI tables, it looks like there are "only" duplicates between the DSDT and one of the SSDT tables. Cheers, Mike
