ACPI ERROR SMBus or IPMI write requires Buff of len 66, found 32...???? linux-3.0.0

Was looking through boot log messages, and this one seemed 'new'...

is this saying that the device requires 66, but the driver allocated 32 bytes?

Or the opposite: i.e. driver requires 66 (according to some new standard**) but the device from 2009, has the 2009 standard of 32?

Just a WAG, but 20110413 looks like a date, that is certainly newer than the system, though it has had firmware upgrades (though none in the past 6 months or so...)...

Seems like a weird thing to be changed across drivers while not providing for some backward compat, but that's purely based on a notion that there's a driver/hw version mismatch now when before, there were no such error messages (maybe the error message is new, and it's always been broken? ;-) )...

as subj says linux-3.0.0, stock vanilla + the latest IMQ networkfilter patch. X64

[   13.722612] ACPI Error: SMBus or IPMI write requires Buffer of length 66, found length 32 (20110413/exfield-285)
[   13.733078] ACPI Error: Method parse/execution failed [\_SB_.PMI0._GHL] (Node ffff880616962b18), AE_AML_BUFFER_LIMIT (20110413/psparse-536)
[   13.746189] ACPI Error: Method parse/execution failed [\_SB_.PMI0._PMC] (Node ffff880616962a78), AE_AML_BUFFER_LIMIT
[   13.755322] dca service started, version 1.12.1[   13.761822]  (20110413/psparse-536)
[   13.765509] ACPI Exception: AE_AML_BUFFER_LIMIT, Evaluating _PMC (20110413/power_meter-773)

Did something that would cause this? I haven't determined what the impact of the above problem is -- it might be some status progs no longer work OR maybe never worked, and now
may have a clue... ? Any insights?

[+cc linux-acpi]

Can you confirm that this error did not occur on a previous version? What is the most recent version where it did not occur?

20110413 is basically the version number of the ACPI CA in Linux. Generally, we don't remove support for old things, so if it worked in the past, it should still work with newer versions.

For completeness, it'd be nice to see the complete dmesg log. In this case, it looks AML-related, so an acpidump would also be useful.

Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

ACPI IPMI support was only added recently, so it's not surprising that it didn't show up previously. The spec is quite clear that IPMI message buffers should be 66 bytes, but a lot of firmware seems to disagree. Not sure what the best thing to do here is.

Matthew Garrett | [email protected]