Most recent kernel where this bug did not occur: Distribution: 'acpi_osi=Linux' made a PCI device window available. Hardware Environment: Toshiba A135-s2386 Software Environment: Fedora 8 Problem Description: with 'acpi_osi=Linux' kernel parameter switched on, one of PCI devices window is available. The dmesg shows: without acpi_osi=Linux < PCI: Cannot allocate resource region 7 of bridge 0000:00:06.0 < PCI: Cannot allocate resource region 8 of bridge 0000:00:06.0 < MEM window: disabled. with acpi_osi=Linux > MEM window: c0100000-c01fffff Steps to reproduce: add acpi_osi=Linux kernel parameter.
Created attachment 14049 [details] output of acpidump acpidump is not avaliable in fedora 8, and utilities links at http://www.lesswatts.org/projects/acpi/utilities.php seems was broken, so I got one from http://people.freebsd.org/~takawata/pacpidump
Created attachment 14050 [details] dmesg with acpi_osi=linux kernel parameter
Created attachment 14051 [details] dmesg without acpi_osi=Linux
Created attachment 14052 [details] dmidecode
Created attachment 14053 [details] lspci -nn output
Hi, Jason, please attach the acpidump binary file instead. You can find the latest pmtools source code at http://www.kernel.org/pub/linux/kernel/people/lenb/acpi/utils/
Created attachment 14138 [details] #./acpidump > acpidump.txt output by pmtools-20071116
Besides the PCI dmesg diff above, do you notice any functional difference with and without acpi_osi=Linux? Device (PCI0) { Name (_HID, EisaId ("PNP0A03")) Name (_ADR, 0x00) Name (_UID, 0x01) Name (_BBN, 0x00) Method (_INI, 0, NotSerialized) { If (CondRefOf (\_OSI, Local0)) { If (\_OSI ("Linux")) { Store (0x00, ECDY) } } ECDY is accessed in multiple places in AML, most notably in the Embedded Controller interface, and the battery code. Do you notice any difference in the battery info or battery function with and without acpi_osi=Linux?
I notised that this laptop often lose its wlan device (from lspci) sometimes when I do cold-boot or wakeup back from sleep(suspend to ram). And the lost PCI window is right the wlan device. So I think this Bug does not belongs to OS but BIOS of the laptop or hardware. Now I can confirm that the 'acpi_osi=Linux' has nothing to do with the lost of the wlan PCI device window. I believe This is NOT a bug of linux ACPI also because the pci window can also get lost when I using Windows. If what I believe is true, I will close the bug report. About the battery function, there is nothing differences when with or without acpi_osi=Linux.
Close this bug because there is no difference w/ and w/o "acpi_osi=linux".