FW: ACPI Warning: Invalid length for Pm1aControlBlock: 32, using default 16 (20090903/tbfadt-607)

FW: ACPI Warning: Invalid length for Pm1aControlBlock: 32, using default 16 (20090903/tbfadt-607)

[Berahalli, Sreenivasa Reddy]

Resending because delivery failure in first attempt.

Sreenivas

_____________________________________________
From: Berahalli, Sreenivasa Reddy
Sent: Friday, January 14, 2011 11:28 AM
To: 'lenb@kernel.org'; 'ming.m.lin@intel.com'; 'robert.moore@intel.com'; 'linux-acpi@vger.kernel.org'; 'linux-kernel@vger.kernel.org'
Cc: K, Narmadha
Subject: ACPI Warning: Invalid length for Pm1aControlBlock: 32, using default 16 (20090903/tbfadt-607)


Hi,

>From kernel version 2.6.31.13, the "acpi_tb_setup_fadt_registers" function is added to check for any incorrect length.
In ACPI tables, this function checks if the length of register and the default value are the same, and will replace if the values are different.
 acpi_tb_setup_fadt_registers() function in /drivers/acpi/acpica/tbfadt.c verifies the length of Pm1aControlBlock.

"ACPI Warning: Invalid length for Pm1aControlBlock: 32, using default 16 (20090903/tbfadt-607)"

This warning message means the length of register in the ACPI tables and the default values are different.
We see it as  over-aggressive check in Linux, with no issue.

Warning messages in dmesg are alarm to the end users.
This issue can be addressed by moving these messages to debug level instead of throwing it by default to dmesg.

Thanks
Sreenivas

Re: FW: ACPI Warning: Invalid length for Pm1aControlBlock: 32, using default 16 (20090903/tbfadt-607)

[Matthew Garrett]

On Fri, Jan 14, 2011 at 06:33:13PM +0000, Berahalli, Sreenivasa Reddy wrote:

> This warning message means the length of register in the ACPI tables and the default values are different.
> We see it as  over-aggressive check in Linux, with no issue.

Linux sees it as a deviation between the BIOS and the spec, which may 
cause issues if it turns out that we treat this behaviour differently to 
tested operating systems. Why not fix your BIOS?

-- 
Matthew Garrett | mjg59@srcf.ucam.org