Since I haven't found any solution to the problem yet (searched the forums too), I'm just wondering if anyone else here had the same problem and/or whether people at Intel are aware of this (and are planning a fix for this issue).
I've bought a D945GCLF2 board some time ago, but I've decided to plug a NIC into the PCI slot quite recently. After booting the system (linux, vanilla 2.6.33 kernel at the moment) and working for a while I get a lot of reports:
Mar 14 03:04:55 atom kernel: [ 7398.965095] ACPI Exception: AE_AML_INFINITE_LOOP, while evaluating GPE method [_L00] (20090903/evgpe-568)
Mar 14 03:04:57 atom kernel: [ 7401.183183] ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.PCI0.LPC_.SMBR] (Node ffff88007f0118a0), AE_AML_INFINITE_LOOP
Mar 14 03:04:57 atom kernel: [ 7401.183301] ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.PCI0.LPC_.INIT] (Node ffff88007f011880), AE_AML_INFINITE_LOOP
Mar 14 03:04:57 atom kernel: [ 7401.183416] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L00] (Node ffff88007f006540), AE_AML_INFINITE_LOOP
According to this linux kernel bugzilla report, the problem is caused by the buggy BIOS, so I thought flashing it with the most recent version (LF94510J.86A.0229.2009.0729.0209 released 07/29/2009 as dmidecode reports) would solve it, but - sadly - it didn't. Right now I'm quite lost, not sure what else I can do (disabling ACPI is not the solution I'm looking for). Does anyone here have any experience with this? Is Intel planning on fixing that anywhere in the future?
Thank you in advance.
As a workaround I'm using my own DSDT table at the moment (changed SMBus' IOport from 0x2000 to 0x3000 - that's where it is when I attach another NIC in the PCI slot - according to lspci). For now it should work, but that's just a workaround...